* [PATCH net v2 3/4] net: ethernet: ti: davinci_emac: fix device reference leak
From: Johan Hovold @ 2016-11-03 17:40 UTC (permalink / raw)
To: Florian Fainelli, Mugunthan V N, Yisen Zhuang, Salil Mehta,
David S. Miller
Cc: netdev, linux-kernel, Johan Hovold, Grygorii Strashko, linux-omap
In-Reply-To: <1478194822-29545-1-git-send-email-johan@kernel.org>
Make sure to drop the references taken by bus_find_device() before
returning from emac_dev_open().
Note that phy_connect still takes a reference to the phy device.
Fixes: 5d69e0076a72 ("net: davinci_emac: switch to new mdio")
Cc: Mugunthan V N <mugunthanvnm@ti.com>
Cc: Grygorii Strashko <grygorii.strashko@ti.com>
Cc: linux-omap@vger.kernel.org
Signed-off-by: Johan Hovold <johan@kernel.org>
---
drivers/net/ethernet/ti/davinci_emac.c | 10 ++++++----
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/ti/davinci_emac.c b/drivers/net/ethernet/ti/davinci_emac.c
index 2fd94a5bc1f3..84fbe5714f8b 100644
--- a/drivers/net/ethernet/ti/davinci_emac.c
+++ b/drivers/net/ethernet/ti/davinci_emac.c
@@ -1410,6 +1410,7 @@ static int emac_dev_open(struct net_device *ndev)
int i = 0;
struct emac_priv *priv = netdev_priv(ndev);
struct phy_device *phydev = NULL;
+ struct device *phy = NULL;
ret = pm_runtime_get_sync(&priv->pdev->dev);
if (ret < 0) {
@@ -1488,19 +1489,20 @@ static int emac_dev_open(struct net_device *ndev)
/* use the first phy on the bus if pdata did not give us a phy id */
if (!phydev && !priv->phy_id) {
- struct device *phy;
-
phy = bus_find_device(&mdio_bus_type, NULL, NULL,
match_first_device);
- if (phy)
+ if (phy) {
priv->phy_id = dev_name(phy);
+ if (!priv->phy_id || !*priv->phy_id)
+ put_device(phy);
+ }
}
if (!phydev && priv->phy_id && *priv->phy_id) {
phydev = phy_connect(ndev, priv->phy_id,
&emac_adjust_link,
PHY_INTERFACE_MODE_MII);
-
+ put_device(phy); /* reference taken by bus_find_device */
if (IS_ERR(phydev)) {
dev_err(emac_dev, "could not connect to phy %s\n",
priv->phy_id);
--
2.7.3
^ permalink raw reply related
* [PATCH net v2 2/4] net: ethernet: ti: cpsw: fix device and of_node leaks
From: Johan Hovold @ 2016-11-03 17:40 UTC (permalink / raw)
To: Florian Fainelli, Mugunthan V N, Yisen Zhuang, Salil Mehta,
David S. Miller
Cc: netdev, linux-kernel, Johan Hovold, Grygorii Strashko, linux-omap
In-Reply-To: <1478194822-29545-1-git-send-email-johan@kernel.org>
Make sure to drop the references taken by of_get_child_by_name() and
bus_find_device() before returning from cpsw_phy_sel().
Note that holding a reference to the cpsw-phy-sel device does not
prevent the devres-managed private data from going away.
Fixes: 5892cd135e16 ("drivers: net: cpsw-phy-sel: Add new driver...")
Cc: Mugunthan V N <mugunthanvnm@ti.com>
Cc: Grygorii Strashko <grygorii.strashko@ti.com>
Cc: linux-omap@vger.kernel.org
Signed-off-by: Johan Hovold <johan@kernel.org>
---
drivers/net/ethernet/ti/cpsw-phy-sel.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/ti/cpsw-phy-sel.c b/drivers/net/ethernet/ti/cpsw-phy-sel.c
index 054a8dd23dae..ba1e45ff6aae 100644
--- a/drivers/net/ethernet/ti/cpsw-phy-sel.c
+++ b/drivers/net/ethernet/ti/cpsw-phy-sel.c
@@ -176,9 +176,12 @@ void cpsw_phy_sel(struct device *dev, phy_interface_t phy_mode, int slave)
}
dev = bus_find_device(&platform_bus_type, NULL, node, match);
+ of_node_put(node);
priv = dev_get_drvdata(dev);
priv->cpsw_phy_sel(priv, phy_mode, slave);
+
+ put_device(dev);
}
EXPORT_SYMBOL_GPL(cpsw_phy_sel);
--
2.7.3
^ permalink raw reply related
* [PATCH net v2 1/4] phy: fix device reference leaks
From: Johan Hovold @ 2016-11-03 17:40 UTC (permalink / raw)
To: Florian Fainelli, Mugunthan V N, Yisen Zhuang, Salil Mehta,
David S. Miller
Cc: netdev, linux-kernel, Johan Hovold
In-Reply-To: <1478194822-29545-1-git-send-email-johan@kernel.org>
Make sure to drop the reference taken by bus_find_device_by_name()
before returning from phy_connect() and phy_attach().
Note that both function still take a reference to the phy device
through phy_attach_direct().
Fixes: e13934563db0 ("[PATCH] PHY Layer fixup")
Cc: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
---
drivers/net/phy/phy_device.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
index e977ba931878..1a4bf8acad78 100644
--- a/drivers/net/phy/phy_device.c
+++ b/drivers/net/phy/phy_device.c
@@ -723,6 +723,7 @@ struct phy_device *phy_connect(struct net_device *dev, const char *bus_id,
phydev = to_phy_device(d);
rc = phy_connect_direct(dev, phydev, handler, interface);
+ put_device(d);
if (rc)
return ERR_PTR(rc);
@@ -953,6 +954,7 @@ struct phy_device *phy_attach(struct net_device *dev, const char *bus_id,
phydev = to_phy_device(d);
rc = phy_attach_direct(dev, phydev, phydev->dev_flags, interface);
+ put_device(d);
if (rc)
return ERR_PTR(rc);
--
2.7.3
^ permalink raw reply related
* [PATCH net v2 0/4] net: fix device reference leaks
From: Johan Hovold @ 2016-11-03 17:40 UTC (permalink / raw)
To: Florian Fainelli, Mugunthan V N, Yisen Zhuang, Salil Mehta,
David S. Miller
Cc: netdev, linux-kernel, Johan Hovold
This series fixes a number of device reference leaks (and one of_node
leak) due to failure to drop the references taken by bus_find_device()
and friends.
Note that the final two patches have been compile tested only.
Thanks,
Johan
v2
- hold reference to cpsw-phy-sel device while accessing private data as
requested by David. Also update the commit message. (patch 1/4)
- add linux-omap on CC where appropriate
Johan Hovold (4):
phy: fix device reference leaks
net: ethernet: ti: cpsw: fix device and of_node leaks
net: ethernet: ti: davinci_emac: fix device reference leak
net: hns: fix device reference leaks
drivers/net/ethernet/hisilicon/hns/hnae.c | 8 +++++++-
drivers/net/ethernet/ti/cpsw-phy-sel.c | 3 +++
drivers/net/ethernet/ti/davinci_emac.c | 10 ++++++----
drivers/net/phy/phy_device.c | 2 ++
4 files changed, 18 insertions(+), 5 deletions(-)
--
2.7.3
^ permalink raw reply
* Re: [PATCH net] net: Check for fullsock in sock_i_uid()
From: Lorenzo Colitti @ 2016-11-03 17:34 UTC (permalink / raw)
To: Eric Dumazet
Cc: Subash Abhinov Kasiviswanathan, netdev@vger.kernel.org,
Eric Dumazet
In-Reply-To: <1478107107.7065.379.camel@edumazet-glaptop3.roam.corp.google.com>
On Thu, Nov 3, 2016 at 2:18 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> Lorenzo, have'nt you already fixed all these bugs ?
Not yet. There's still a fair bit of out-of-tree code left. Other than
per-UID routing, xt_qtaguid is the big one, but there's also xt_quota2
and xt_idletimer. to fix.
> if (skb && skb->sk)
> timer->uid = from_kuid_munged(current_user_ns(),
> sock_i_uid(skb_to_full_sk(skb)));
Thanks! I didn't know about skb_to_full_sk.
^ permalink raw reply
* [PATCH net-next v2 3/3] net: inet: Support UID-based routing in IP protocols.
From: Lorenzo Colitti @ 2016-11-03 17:23 UTC (permalink / raw)
To: netdev; +Cc: davem, ek, eric.dumazet, zenczykowski, Lorenzo Colitti
In-Reply-To: <1478193823-60516-1-git-send-email-lorenzo@google.com>
- Use the UID in routing lookups made by protocol connect() and
sendmsg() functions.
- Make sure that routing lookups triggered by incoming packets
(e.g., Path MTU discovery) take the UID of the socket into
account.
- For packets not associated with a userspace socket, (e.g., ping
replies) use UID 0 inside the user namespace corresponding to
the network namespace the socket belongs to. This allows
all namespaces to apply routing and iptables rules to
kernel-originated traffic in that namespaces by matching UID 0.
This is better than using the UID of the kernel socket that is
sending the traffic, because the UID of kernel sockets created
at namespace creation time (e.g., the per-processor ICMP and
TCP sockets) is the UID of the user that created the socket,
which might not be mapped in the namespace.
Tested: compiles allnoconfig, allyesconfig, allmodconfig
Tested: https://android-review.googlesource.com/253302
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
---
include/net/flow.h | 4 +++-
include/net/ip.h | 1 +
include/net/ip6_route.h | 5 +++--
include/net/route.h | 5 +++--
net/ipv4/icmp.c | 2 ++
net/ipv4/inet_connection_sock.c | 4 ++--
net/ipv4/ip_output.c | 3 ++-
net/ipv4/ping.c | 3 ++-
net/ipv4/raw.c | 2 +-
net/ipv4/route.c | 26 +++++++++++++++-----------
net/ipv4/syncookies.c | 2 +-
net/ipv4/tcp_ipv4.c | 9 ++++++---
net/ipv4/udp.c | 3 ++-
net/ipv6/af_inet6.c | 1 +
net/ipv6/ah6.c | 5 +++--
net/ipv6/datagram.c | 1 +
net/ipv6/esp6.c | 5 +++--
net/ipv6/icmp.c | 7 +++++--
net/ipv6/inet6_connection_sock.c | 2 ++
net/ipv6/ip6_gre.c | 4 ++++
net/ipv6/ip6_tunnel.c | 4 ++++
net/ipv6/ip6_vti.c | 5 +++--
net/ipv6/ipcomp6.c | 5 +++--
net/ipv6/netfilter.c | 1 +
net/ipv6/ping.c | 1 +
net/ipv6/raw.c | 1 +
net/ipv6/route.c | 13 +++++++++----
net/ipv6/syncookies.c | 1 +
net/ipv6/tcp_ipv6.c | 2 ++
net/ipv6/udp.c | 1 +
net/l2tp/l2tp_ip6.c | 1 +
31 files changed, 89 insertions(+), 40 deletions(-)
diff --git a/include/net/flow.h b/include/net/flow.h
index 51373f3..6bbbca8 100644
--- a/include/net/flow.h
+++ b/include/net/flow.h
@@ -96,7 +96,8 @@ static inline void flowi4_init_output(struct flowi4 *fl4, int oif,
__u32 mark, __u8 tos, __u8 scope,
__u8 proto, __u8 flags,
__be32 daddr, __be32 saddr,
- __be16 dport, __be16 sport)
+ __be16 dport, __be16 sport,
+ kuid_t uid)
{
fl4->flowi4_oif = oif;
fl4->flowi4_iif = LOOPBACK_IFINDEX;
@@ -107,6 +108,7 @@ static inline void flowi4_init_output(struct flowi4 *fl4, int oif,
fl4->flowi4_flags = flags;
fl4->flowi4_secid = 0;
fl4->flowi4_tun_key.tun_id = 0;
+ fl4->flowi4_uid = uid;
fl4->daddr = daddr;
fl4->saddr = saddr;
fl4->fl4_dport = dport;
diff --git a/include/net/ip.h b/include/net/ip.h
index 5413883..55cdaac 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -179,6 +179,7 @@ struct ip_reply_arg {
/* -1 if not needed */
int bound_dev_if;
u8 tos;
+ kuid_t uid;
};
#define IP_REPLY_ARG_NOSRCCHECK 1
diff --git a/include/net/ip6_route.h b/include/net/ip6_route.h
index f83e78d..9dc2c18 100644
--- a/include/net/ip6_route.h
+++ b/include/net/ip6_route.h
@@ -140,9 +140,10 @@ int rt6_route_rcv(struct net_device *dev, u8 *opt, int len,
const struct in6_addr *gwaddr);
void ip6_update_pmtu(struct sk_buff *skb, struct net *net, __be32 mtu, int oif,
- u32 mark);
+ u32 mark, kuid_t uid);
void ip6_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, __be32 mtu);
-void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark);
+void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark,
+ kuid_t uid);
void ip6_redirect_no_header(struct sk_buff *skb, struct net *net, int oif,
u32 mark);
void ip6_sk_redirect(struct sk_buff *skb, struct sock *sk);
diff --git a/include/net/route.h b/include/net/route.h
index 0429d47..c0874c8 100644
--- a/include/net/route.h
+++ b/include/net/route.h
@@ -153,7 +153,7 @@ static inline struct rtable *ip_route_output_ports(struct net *net, struct flowi
flowi4_init_output(fl4, oif, sk ? sk->sk_mark : 0, tos,
RT_SCOPE_UNIVERSE, proto,
sk ? inet_sk_flowi_flags(sk) : 0,
- daddr, saddr, dport, sport);
+ daddr, saddr, dport, sport, sock_net_uid(net, sk));
if (sk)
security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
return ip_route_output_flow(net, fl4, sk);
@@ -269,7 +269,8 @@ static inline void ip_route_connect_init(struct flowi4 *fl4, __be32 dst, __be32
flow_flags |= FLOWI_FLAG_ANYSRC;
flowi4_init_output(fl4, oif, sk->sk_mark, tos, RT_SCOPE_UNIVERSE,
- protocol, flow_flags, dst, src, dport, sport);
+ protocol, flow_flags, dst, src, dport, sport,
+ sk->sk_uid);
}
static inline struct rtable *ip_route_connect(struct flowi4 *fl4,
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 38abe70..53a890b 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -425,6 +425,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
fl4.daddr = daddr;
fl4.saddr = saddr;
fl4.flowi4_mark = mark;
+ fl4.flowi4_uid = sock_net_uid(net, NULL);
fl4.flowi4_tos = RT_TOS(ip_hdr(skb)->tos);
fl4.flowi4_proto = IPPROTO_ICMP;
fl4.flowi4_oif = l3mdev_master_ifindex(skb->dev);
@@ -473,6 +474,7 @@ static struct rtable *icmp_route_lookup(struct net *net,
param->replyopts.opt.opt.faddr : iph->saddr);
fl4->saddr = saddr;
fl4->flowi4_mark = mark;
+ fl4->flowi4_uid = sock_net_uid(net, NULL);
fl4->flowi4_tos = RT_TOS(tos);
fl4->flowi4_proto = IPPROTO_ICMP;
fl4->fl4_icmp_type = type;
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 61a9dee..d5d3ead 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -415,7 +415,7 @@ struct dst_entry *inet_csk_route_req(const struct sock *sk,
sk->sk_protocol, inet_sk_flowi_flags(sk),
(opt && opt->opt.srr) ? opt->opt.faddr : ireq->ir_rmt_addr,
ireq->ir_loc_addr, ireq->ir_rmt_port,
- htons(ireq->ir_num));
+ htons(ireq->ir_num), sk->sk_uid);
security_req_classify_flow(req, flowi4_to_flowi(fl4));
rt = ip_route_output_flow(net, fl4, sk);
if (IS_ERR(rt))
@@ -452,7 +452,7 @@ struct dst_entry *inet_csk_route_child_sock(const struct sock *sk,
sk->sk_protocol, inet_sk_flowi_flags(sk),
(opt && opt->opt.srr) ? opt->opt.faddr : ireq->ir_rmt_addr,
ireq->ir_loc_addr, ireq->ir_rmt_port,
- htons(ireq->ir_num));
+ htons(ireq->ir_num), sk->sk_uid);
security_req_classify_flow(req, flowi4_to_flowi(fl4));
rt = ip_route_output_flow(net, fl4, sk);
if (IS_ERR(rt))
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 03e7f73..37dfacd3 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -1587,7 +1587,8 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb,
RT_SCOPE_UNIVERSE, ip_hdr(skb)->protocol,
ip_reply_arg_flowi_flags(arg),
daddr, saddr,
- tcp_hdr(skb)->source, tcp_hdr(skb)->dest);
+ tcp_hdr(skb)->source, tcp_hdr(skb)->dest,
+ arg->uid);
security_skb_classify_flow(skb, flowi4_to_flowi(&fl4));
rt = ip_route_output_key(net, &fl4);
if (IS_ERR(rt))
diff --git a/net/ipv4/ping.c b/net/ipv4/ping.c
index 205e200..d11129f 100644
--- a/net/ipv4/ping.c
+++ b/net/ipv4/ping.c
@@ -789,7 +789,8 @@ static int ping_v4_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
flowi4_init_output(&fl4, ipc.oif, sk->sk_mark, tos,
RT_SCOPE_UNIVERSE, sk->sk_protocol,
- inet_sk_flowi_flags(sk), faddr, saddr, 0, 0);
+ inet_sk_flowi_flags(sk), faddr, saddr, 0, 0,
+ sk->sk_uid);
security_sk_classify_flow(sk, flowi4_to_flowi(&fl4));
rt = ip_route_output_flow(net, &fl4, sk);
diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c
index 6a0bd68..838ea5e 100644
--- a/net/ipv4/raw.c
+++ b/net/ipv4/raw.c
@@ -606,7 +606,7 @@ static int raw_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
inet->hdrincl ? IPPROTO_RAW : sk->sk_protocol,
inet_sk_flowi_flags(sk) |
(inet->hdrincl ? FLOWI_FLAG_KNOWN_NH : 0),
- daddr, saddr, 0, 0);
+ daddr, saddr, 0, 0, sk->sk_uid);
if (!inet->hdrincl) {
rfv.msg = msg;
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 92e59a6..2355883 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -507,7 +507,8 @@ void __ip_select_ident(struct net *net, struct iphdr *iph, int segs)
}
EXPORT_SYMBOL(__ip_select_ident);
-static void __build_flow_key(struct flowi4 *fl4, const struct sock *sk,
+static void __build_flow_key(const struct net *net, struct flowi4 *fl4,
+ const struct sock *sk,
const struct iphdr *iph,
int oif, u8 tos,
u8 prot, u32 mark, int flow_flags)
@@ -523,7 +524,8 @@ static void __build_flow_key(struct flowi4 *fl4, const struct sock *sk,
flowi4_init_output(fl4, oif, mark, tos,
RT_SCOPE_UNIVERSE, prot,
flow_flags,
- iph->daddr, iph->saddr, 0, 0);
+ iph->daddr, iph->saddr, 0, 0,
+ sock_net_uid(net, sk));
}
static void build_skb_flow_key(struct flowi4 *fl4, const struct sk_buff *skb,
@@ -535,7 +537,7 @@ static void build_skb_flow_key(struct flowi4 *fl4, const struct sk_buff *skb,
u8 prot = iph->protocol;
u32 mark = skb->mark;
- __build_flow_key(fl4, sk, iph, oif, tos, prot, mark, 0);
+ __build_flow_key(sock_net(sk), fl4, sk, iph, oif, tos, prot, mark, 0);
}
static void build_sk_flow_key(struct flowi4 *fl4, const struct sock *sk)
@@ -552,7 +554,7 @@ static void build_sk_flow_key(struct flowi4 *fl4, const struct sock *sk)
RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE,
inet->hdrincl ? IPPROTO_RAW : sk->sk_protocol,
inet_sk_flowi_flags(sk),
- daddr, inet->inet_saddr, 0, 0);
+ daddr, inet->inet_saddr, 0, 0, sk->sk_uid);
rcu_read_unlock();
}
@@ -800,7 +802,7 @@ static void ip_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_buf
rt = (struct rtable *) dst;
- __build_flow_key(&fl4, sk, iph, oif, tos, prot, mark, 0);
+ __build_flow_key(sock_net(sk), &fl4, sk, iph, oif, tos, prot, mark, 0);
__ip_do_redirect(rt, skb, &fl4, true);
}
@@ -1018,7 +1020,7 @@ void ipv4_update_pmtu(struct sk_buff *skb, struct net *net, u32 mtu,
if (!mark)
mark = IP4_REPLY_MARK(net, skb->mark);
- __build_flow_key(&fl4, NULL, iph, oif,
+ __build_flow_key(net, &fl4, NULL, iph, oif,
RT_TOS(iph->tos), protocol, mark, flow_flags);
rt = __ip_route_output_key(net, &fl4);
if (!IS_ERR(rt)) {
@@ -1034,7 +1036,7 @@ static void __ipv4_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, u32 mtu)
struct flowi4 fl4;
struct rtable *rt;
- __build_flow_key(&fl4, sk, iph, 0, 0, 0, 0, 0);
+ __build_flow_key(sock_net(sk), &fl4, sk, iph, 0, 0, 0, 0, 0);
if (!fl4.flowi4_mark)
fl4.flowi4_mark = IP4_REPLY_MARK(sock_net(sk), skb->mark);
@@ -1053,6 +1055,7 @@ void ipv4_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, u32 mtu)
struct rtable *rt;
struct dst_entry *odst = NULL;
bool new = false;
+ struct net *net = sock_net(sk);
bh_lock_sock(sk);
@@ -1066,7 +1069,7 @@ void ipv4_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, u32 mtu)
goto out;
}
- __build_flow_key(&fl4, sk, iph, 0, 0, 0, 0, 0);
+ __build_flow_key(net, &fl4, sk, iph, 0, 0, 0, 0, 0);
rt = (struct rtable *)odst;
if (odst->obsolete && !odst->ops->check(odst, 0)) {
@@ -1106,7 +1109,7 @@ void ipv4_redirect(struct sk_buff *skb, struct net *net,
struct flowi4 fl4;
struct rtable *rt;
- __build_flow_key(&fl4, NULL, iph, oif,
+ __build_flow_key(net, &fl4, NULL, iph, oif,
RT_TOS(iph->tos), protocol, mark, flow_flags);
rt = __ip_route_output_key(net, &fl4);
if (!IS_ERR(rt)) {
@@ -1121,9 +1124,10 @@ void ipv4_sk_redirect(struct sk_buff *skb, struct sock *sk)
const struct iphdr *iph = (const struct iphdr *) skb->data;
struct flowi4 fl4;
struct rtable *rt;
+ struct net *net = sock_net(sk);
- __build_flow_key(&fl4, sk, iph, 0, 0, 0, 0, 0);
- rt = __ip_route_output_key(sock_net(sk), &fl4);
+ __build_flow_key(net, &fl4, sk, iph, 0, 0, 0, 0, 0);
+ rt = __ip_route_output_key(net, &fl4);
if (!IS_ERR(rt)) {
__ip_do_redirect(rt, skb, &fl4, false);
ip_rt_put(rt);
diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index e3c4043..0dc6286 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -372,7 +372,7 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb)
RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE, IPPROTO_TCP,
inet_sk_flowi_flags(sk),
opt->srr ? opt->faddr : ireq->ir_rmt_addr,
- ireq->ir_loc_addr, th->source, th->dest);
+ ireq->ir_loc_addr, th->source, th->dest, sk->sk_uid);
security_req_classify_flow(req, flowi4_to_flowi(&fl4));
rt = ip_route_output_key(sock_net(sk), &fl4);
if (IS_ERR(rt)) {
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index b9b82826..6491b7c 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -691,6 +691,7 @@ static void tcp_v4_send_reset(const struct sock *sk, struct sk_buff *skb)
offsetof(struct inet_timewait_sock, tw_bound_dev_if));
arg.tos = ip_hdr(skb)->tos;
+ arg.uid = sock_net_uid(net, sk && sk_fullsock(sk) ? sk : NULL);
local_bh_disable();
ip_send_unicast_reply(*this_cpu_ptr(net->ipv4.tcp_sk),
skb, &TCP_SKB_CB(skb)->header.h4.opt,
@@ -711,7 +712,7 @@ static void tcp_v4_send_reset(const struct sock *sk, struct sk_buff *skb)
outside socket context is ugly, certainly. What can I do?
*/
-static void tcp_v4_send_ack(struct net *net,
+static void tcp_v4_send_ack(const struct sock *sk,
struct sk_buff *skb, u32 seq, u32 ack,
u32 win, u32 tsval, u32 tsecr, int oif,
struct tcp_md5sig_key *key,
@@ -726,6 +727,7 @@ static void tcp_v4_send_ack(struct net *net,
#endif
];
} rep;
+ struct net *net = sock_net(sk);
struct ip_reply_arg arg;
memset(&rep.th, 0, sizeof(struct tcphdr));
@@ -775,6 +777,7 @@ static void tcp_v4_send_ack(struct net *net,
if (oif)
arg.bound_dev_if = oif;
arg.tos = tos;
+ arg.uid = sock_net_uid(net, sk_fullsock(sk) ? sk : NULL);
local_bh_disable();
ip_send_unicast_reply(*this_cpu_ptr(net->ipv4.tcp_sk),
skb, &TCP_SKB_CB(skb)->header.h4.opt,
@@ -790,7 +793,7 @@ static void tcp_v4_timewait_ack(struct sock *sk, struct sk_buff *skb)
struct inet_timewait_sock *tw = inet_twsk(sk);
struct tcp_timewait_sock *tcptw = tcp_twsk(sk);
- tcp_v4_send_ack(sock_net(sk), skb,
+ tcp_v4_send_ack(sk, skb,
tcptw->tw_snd_nxt, tcptw->tw_rcv_nxt,
tcptw->tw_rcv_wnd >> tw->tw_rcv_wscale,
tcp_time_stamp + tcptw->tw_ts_offset,
@@ -818,7 +821,7 @@ static void tcp_v4_reqsk_send_ack(const struct sock *sk, struct sk_buff *skb,
* exception of <SYN> segments, MUST be right-shifted by
* Rcv.Wind.Shift bits:
*/
- tcp_v4_send_ack(sock_net(sk), skb, seq,
+ tcp_v4_send_ack(sk, skb, seq,
tcp_rsk(req)->rcv_nxt,
req->rsk_rcv_wnd >> inet_rsk(req)->rcv_wscale,
tcp_time_stamp,
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index 195992e..a32a890 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -1019,7 +1019,8 @@ int udp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
flowi4_init_output(fl4, ipc.oif, sk->sk_mark, tos,
RT_SCOPE_UNIVERSE, sk->sk_protocol,
flow_flags,
- faddr, saddr, dport, inet->inet_sport);
+ faddr, saddr, dport, inet->inet_sport,
+ sk->sk_uid);
security_sk_classify_flow(sk, flowi4_to_flowi(fl4));
rt = ip_route_output_flow(net, fl4, sk);
diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
index 46ad699..c86911b 100644
--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -678,6 +678,7 @@ int inet6_sk_rebuild_header(struct sock *sk)
fl6.flowi6_mark = sk->sk_mark;
fl6.fl6_dport = inet->inet_dport;
fl6.fl6_sport = inet->inet_sport;
+ fl6.flowi6_uid = sk->sk_uid;
security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
rcu_read_lock();
diff --git a/net/ipv6/ah6.c b/net/ipv6/ah6.c
index 0630a4d5..189eb10 100644
--- a/net/ipv6/ah6.c
+++ b/net/ipv6/ah6.c
@@ -662,9 +662,10 @@ static int ah6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
return 0;
if (type == NDISC_REDIRECT)
- ip6_redirect(skb, net, skb->dev->ifindex, 0);
+ ip6_redirect(skb, net, skb->dev->ifindex, 0,
+ sock_net_uid(net, NULL));
else
- ip6_update_pmtu(skb, net, info, 0, 0);
+ ip6_update_pmtu(skb, net, info, 0, 0, sock_net_uid(net, NULL));
xfrm_state_put(x);
return 0;
diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c
index 37874e2..6d8a66a 100644
--- a/net/ipv6/datagram.c
+++ b/net/ipv6/datagram.c
@@ -54,6 +54,7 @@ static void ip6_datagram_flow_key_init(struct flowi6 *fl6, struct sock *sk)
fl6->fl6_dport = inet->inet_dport;
fl6->fl6_sport = inet->inet_sport;
fl6->flowlabel = np->flow_label;
+ fl6->flowi6_uid = sk->sk_uid;
if (!fl6->flowi6_oif)
fl6->flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
diff --git a/net/ipv6/esp6.c b/net/ipv6/esp6.c
index 060a60b..218f0cb 100644
--- a/net/ipv6/esp6.c
+++ b/net/ipv6/esp6.c
@@ -474,9 +474,10 @@ static int esp6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
return 0;
if (type == NDISC_REDIRECT)
- ip6_redirect(skb, net, skb->dev->ifindex, 0);
+ ip6_redirect(skb, net, skb->dev->ifindex, 0,
+ sock_net_uid(net, NULL));
else
- ip6_update_pmtu(skb, net, info, 0, 0);
+ ip6_update_pmtu(skb, net, info, 0, 0, sock_net_uid(net, NULL));
xfrm_state_put(x);
return 0;
diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
index bd59c34..ab249fe 100644
--- a/net/ipv6/icmp.c
+++ b/net/ipv6/icmp.c
@@ -92,9 +92,10 @@ static void icmpv6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
struct net *net = dev_net(skb->dev);
if (type == ICMPV6_PKT_TOOBIG)
- ip6_update_pmtu(skb, net, info, 0, 0);
+ ip6_update_pmtu(skb, net, info, 0, 0, sock_net_uid(net, NULL));
else if (type == NDISC_REDIRECT)
- ip6_redirect(skb, net, skb->dev->ifindex, 0);
+ ip6_redirect(skb, net, skb->dev->ifindex, 0,
+ sock_net_uid(net, NULL));
if (!(type & ICMPV6_INFOMSG_MASK))
if (icmp6->icmp6_type == ICMPV6_ECHO_REQUEST)
@@ -484,6 +485,7 @@ static void icmp6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info,
fl6.flowi6_oif = iif;
fl6.fl6_icmp_type = type;
fl6.fl6_icmp_code = code;
+ fl6.flowi6_uid = sock_net_uid(net, NULL);
security_skb_classify_flow(skb, flowi6_to_flowi(&fl6));
sk = icmpv6_xmit_lock(net);
@@ -658,6 +660,7 @@ static void icmpv6_echo_reply(struct sk_buff *skb)
fl6.flowi6_oif = skb->dev->ifindex;
fl6.fl6_icmp_type = ICMPV6_ECHO_REPLY;
fl6.flowi6_mark = mark;
+ fl6.flowi6_uid = sock_net_uid(net, NULL);
security_skb_classify_flow(skb, flowi6_to_flowi(&fl6));
sk = icmpv6_xmit_lock(net);
diff --git a/net/ipv6/inet6_connection_sock.c b/net/ipv6/inet6_connection_sock.c
index 532c3ef..1c86c47 100644
--- a/net/ipv6/inet6_connection_sock.c
+++ b/net/ipv6/inet6_connection_sock.c
@@ -88,6 +88,7 @@ struct dst_entry *inet6_csk_route_req(const struct sock *sk,
fl6->flowi6_mark = ireq->ir_mark;
fl6->fl6_dport = ireq->ir_rmt_port;
fl6->fl6_sport = htons(ireq->ir_num);
+ fl6->flowi6_uid = sk->sk_uid;
security_req_classify_flow(req, flowi6_to_flowi(fl6));
dst = ip6_dst_lookup_flow(sk, fl6, final_p);
@@ -136,6 +137,7 @@ static struct dst_entry *inet6_csk_route_socket(struct sock *sk,
fl6->flowi6_mark = sk->sk_mark;
fl6->fl6_sport = inet->inet_sport;
fl6->fl6_dport = inet->inet_dport;
+ fl6->flowi6_uid = sk->sk_uid;
security_sk_classify_flow(sk, flowi6_to_flowi(fl6));
rcu_read_lock();
diff --git a/net/ipv6/ip6_gre.c b/net/ipv6/ip6_gre.c
index d7d6d3a..710bc79 100644
--- a/net/ipv6/ip6_gre.c
+++ b/net/ipv6/ip6_gre.c
@@ -548,6 +548,8 @@ static inline int ip6gre_xmit_ipv4(struct sk_buff *skb, struct net_device *dev)
if (t->parms.flags & IP6_TNL_F_USE_ORIG_FWMARK)
fl6.flowi6_mark = skb->mark;
+ fl6.flowi6_uid = sock_net_uid(dev_net(dev), NULL);
+
err = gre_handle_offloads(skb, !!(t->parms.o_flags & TUNNEL_CSUM));
if (err)
return -1;
@@ -602,6 +604,8 @@ static inline int ip6gre_xmit_ipv6(struct sk_buff *skb, struct net_device *dev)
if (t->parms.flags & IP6_TNL_F_USE_ORIG_FWMARK)
fl6.flowi6_mark = skb->mark;
+ fl6.flowi6_uid = sock_net_uid(dev_net(dev), NULL);
+
if (gre_handle_offloads(skb, !!(t->parms.o_flags & TUNNEL_CSUM)))
return -1;
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index 03e050d..bfa889c 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -1240,6 +1240,8 @@ ip4ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
fl6.flowi6_mark = skb->mark;
}
+ fl6.flowi6_uid = sock_net_uid(dev_net(dev), NULL);
+
if (iptunnel_handle_offloads(skb, SKB_GSO_IPXIP6))
return -1;
@@ -1318,6 +1320,8 @@ ip6ip6_tnl_xmit(struct sk_buff *skb, struct net_device *dev)
fl6.flowi6_mark = skb->mark;
}
+ fl6.flowi6_uid = sock_net_uid(dev_net(dev), NULL);
+
if (iptunnel_handle_offloads(skb, SKB_GSO_IPXIP6))
return -1;
diff --git a/net/ipv6/ip6_vti.c b/net/ipv6/ip6_vti.c
index 35c5b2d..af3f0e0 100644
--- a/net/ipv6/ip6_vti.c
+++ b/net/ipv6/ip6_vti.c
@@ -608,9 +608,10 @@ static int vti6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
return 0;
if (type == NDISC_REDIRECT)
- ip6_redirect(skb, net, skb->dev->ifindex, 0);
+ ip6_redirect(skb, net, skb->dev->ifindex, 0,
+ sock_net_uid(net, NULL));
else
- ip6_update_pmtu(skb, net, info, 0, 0);
+ ip6_update_pmtu(skb, net, info, 0, 0, sock_net_uid(net, NULL));
xfrm_state_put(x);
return 0;
diff --git a/net/ipv6/ipcomp6.c b/net/ipv6/ipcomp6.c
index 1b9316e..54d165b 100644
--- a/net/ipv6/ipcomp6.c
+++ b/net/ipv6/ipcomp6.c
@@ -74,9 +74,10 @@ static int ipcomp6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
return 0;
if (type == NDISC_REDIRECT)
- ip6_redirect(skb, net, skb->dev->ifindex, 0);
+ ip6_redirect(skb, net, skb->dev->ifindex, 0,
+ sock_net_uid(net, NULL));
else
- ip6_update_pmtu(skb, net, info, 0, 0);
+ ip6_update_pmtu(skb, net, info, 0, 0, sock_net_uid(net, NULL));
xfrm_state_put(x);
return 0;
diff --git a/net/ipv6/netfilter.c b/net/ipv6/netfilter.c
index d11c468..39970e2 100644
--- a/net/ipv6/netfilter.c
+++ b/net/ipv6/netfilter.c
@@ -26,6 +26,7 @@ int ip6_route_me_harder(struct net *net, struct sk_buff *skb)
struct flowi6 fl6 = {
.flowi6_oif = skb->sk ? skb->sk->sk_bound_dev_if : 0,
.flowi6_mark = skb->mark,
+ .flowi6_uid = sock_net_uid(net, skb->sk),
.daddr = iph->daddr,
.saddr = iph->saddr,
};
diff --git a/net/ipv6/ping.c b/net/ipv6/ping.c
index 66e2d9d..e1f8b34 100644
--- a/net/ipv6/ping.c
+++ b/net/ipv6/ping.c
@@ -113,6 +113,7 @@ static int ping_v6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
fl6.daddr = *daddr;
fl6.flowi6_oif = oif;
fl6.flowi6_mark = sk->sk_mark;
+ fl6.flowi6_uid = sk->sk_uid;
fl6.fl6_icmp_type = user_icmph.icmp6_type;
fl6.fl6_icmp_code = user_icmph.icmp6_code;
security_sk_classify_flow(sk, flowi6_to_flowi(&fl6));
diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c
index 610e093..291ebc2 100644
--- a/net/ipv6/raw.c
+++ b/net/ipv6/raw.c
@@ -776,6 +776,7 @@ static int rawv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
memset(&fl6, 0, sizeof(fl6));
fl6.flowi6_mark = sk->sk_mark;
+ fl6.flowi6_uid = sk->sk_uid;
ipc6.hlimit = -1;
ipc6.tclass = -1;
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index fdb9c87..6aa014e 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -1405,7 +1405,7 @@ static void ip6_rt_update_pmtu(struct dst_entry *dst, struct sock *sk,
}
void ip6_update_pmtu(struct sk_buff *skb, struct net *net, __be32 mtu,
- int oif, u32 mark)
+ int oif, u32 mark, kuid_t uid)
{
const struct ipv6hdr *iph = (struct ipv6hdr *) skb->data;
struct dst_entry *dst;
@@ -1417,6 +1417,7 @@ void ip6_update_pmtu(struct sk_buff *skb, struct net *net, __be32 mtu,
fl6.daddr = iph->daddr;
fl6.saddr = iph->saddr;
fl6.flowlabel = ip6_flowinfo(iph);
+ fl6.flowi6_uid = uid;
dst = ip6_route_output(net, NULL, &fl6);
if (!dst->error)
@@ -1430,7 +1431,7 @@ void ip6_sk_update_pmtu(struct sk_buff *skb, struct sock *sk, __be32 mtu)
struct dst_entry *dst;
ip6_update_pmtu(skb, sock_net(sk), mtu,
- sk->sk_bound_dev_if, sk->sk_mark);
+ sk->sk_bound_dev_if, sk->sk_mark, sk->sk_uid);
dst = __sk_dst_get(sk);
if (!dst || !dst->obsolete ||
@@ -1522,7 +1523,8 @@ static struct dst_entry *ip6_route_redirect(struct net *net,
flags, __ip6_route_redirect);
}
-void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark)
+void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark,
+ kuid_t uid)
{
const struct ipv6hdr *iph = (struct ipv6hdr *) skb->data;
struct dst_entry *dst;
@@ -1535,6 +1537,7 @@ void ip6_redirect(struct sk_buff *skb, struct net *net, int oif, u32 mark)
fl6.daddr = iph->daddr;
fl6.saddr = iph->saddr;
fl6.flowlabel = ip6_flowinfo(iph);
+ fl6.flowi6_uid = uid;
dst = ip6_route_redirect(net, &fl6, &ipv6_hdr(skb)->saddr);
rt6_do_redirect(dst, NULL, skb);
@@ -1556,6 +1559,7 @@ void ip6_redirect_no_header(struct sk_buff *skb, struct net *net, int oif,
fl6.flowi6_mark = mark;
fl6.daddr = msg->dest;
fl6.saddr = iph->daddr;
+ fl6.flowi6_uid = sock_net_uid(net, NULL);
dst = ip6_route_redirect(net, &fl6, &iph->saddr);
rt6_do_redirect(dst, NULL, skb);
@@ -1564,7 +1568,8 @@ void ip6_redirect_no_header(struct sk_buff *skb, struct net *net, int oif,
void ip6_sk_redirect(struct sk_buff *skb, struct sock *sk)
{
- ip6_redirect(skb, sock_net(sk), sk->sk_bound_dev_if, sk->sk_mark);
+ ip6_redirect(skb, sock_net(sk), sk->sk_bound_dev_if, sk->sk_mark,
+ sk->sk_uid);
}
EXPORT_SYMBOL_GPL(ip6_sk_redirect);
diff --git a/net/ipv6/syncookies.c b/net/ipv6/syncookies.c
index 59c4839..97830a6 100644
--- a/net/ipv6/syncookies.c
+++ b/net/ipv6/syncookies.c
@@ -227,6 +227,7 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
fl6.flowi6_mark = ireq->ir_mark;
fl6.fl6_dport = ireq->ir_rmt_port;
fl6.fl6_sport = inet_sk(sk)->inet_sport;
+ fl6.flowi6_uid = sk->sk_uid;
security_req_classify_flow(req, flowi6_to_flowi(&fl6));
dst = ip6_dst_lookup_flow(sk, &fl6, final_p);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 5a27ab4..aece1b1 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -233,6 +233,7 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
fl6.flowi6_mark = sk->sk_mark;
fl6.fl6_dport = usin->sin6_port;
fl6.fl6_sport = inet->inet_sport;
+ fl6.flowi6_uid = sk->sk_uid;
opt = rcu_dereference_protected(np->opt, lockdep_sock_is_held(sk));
final_p = fl6_update_dst(&fl6, opt, &final);
@@ -824,6 +825,7 @@ static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32
fl6.flowi6_mark = IP6_REPLY_MARK(net, skb->mark);
fl6.fl6_dport = t1->dest;
fl6.fl6_sport = t1->source;
+ fl6.flowi6_uid = sock_net_uid(net, sk && sk_fullsock(sk) ? sk : NULL);
security_skb_classify_flow(skb, flowi6_to_flowi(&fl6));
/* Pass a socket to ip6_dst_lookup either it is for RST
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index a7700bb..9103c5c 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -1138,6 +1138,7 @@ int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
fl6.flowi6_oif = np->sticky_pktinfo.ipi6_ifindex;
fl6.flowi6_mark = sk->sk_mark;
+ fl6.flowi6_uid = sk->sk_uid;
sockc.tsflags = sk->sk_tsflags;
if (msg->msg_controllen) {
diff --git a/net/l2tp/l2tp_ip6.c b/net/l2tp/l2tp_ip6.c
index ad3468c..1cea54f 100644
--- a/net/l2tp/l2tp_ip6.c
+++ b/net/l2tp/l2tp_ip6.c
@@ -519,6 +519,7 @@ static int l2tp_ip6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
memset(&fl6, 0, sizeof(fl6));
fl6.flowi6_mark = sk->sk_mark;
+ fl6.flowi6_uid = sk->sk_uid;
ipc6.hlimit = -1;
ipc6.tclass = -1;
--
2.8.0.rc3.226.g39d4020
^ permalink raw reply related
* [PATCH net-next v2 2/3] net: core: add UID to flows, rules, and routes
From: Lorenzo Colitti @ 2016-11-03 17:23 UTC (permalink / raw)
To: netdev; +Cc: davem, ek, eric.dumazet, zenczykowski, Lorenzo Colitti
In-Reply-To: <1478193823-60516-1-git-send-email-lorenzo@google.com>
- Define a new FIB rule attributes, FRA_UID_RANGE, to describe a
range of UIDs.
- Define a RTA_UID attribute for per-UID route lookups and dumps.
- Support passing these attributes to and from userspace via
rtnetlink. The value INVALID_UID indicates no UID was
specified.
- Add a UID field to the flow structures.
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
---
include/net/fib_rules.h | 9 ++++-
include/net/flow.h | 5 +++
include/uapi/linux/fib_rules.h | 6 ++++
include/uapi/linux/rtnetlink.h | 1 +
net/core/fib_rules.c | 74 ++++++++++++++++++++++++++++++++++++++++--
net/ipv4/fib_frontend.c | 1 +
net/ipv4/route.c | 11 +++++++
net/ipv6/route.c | 7 ++++
8 files changed, 111 insertions(+), 3 deletions(-)
diff --git a/include/net/fib_rules.h b/include/net/fib_rules.h
index 456e4a6..8dbfdf7 100644
--- a/include/net/fib_rules.h
+++ b/include/net/fib_rules.h
@@ -8,6 +8,11 @@
#include <net/flow.h>
#include <net/rtnetlink.h>
+struct fib_kuid_range {
+ kuid_t start;
+ kuid_t end;
+};
+
struct fib_rule {
struct list_head list;
int iifindex;
@@ -30,6 +35,7 @@ struct fib_rule {
int suppress_prefixlen;
char iifname[IFNAMSIZ];
char oifname[IFNAMSIZ];
+ struct fib_kuid_range uid_range;
struct rcu_head rcu;
};
@@ -92,7 +98,8 @@ struct fib_rules_ops {
[FRA_SUPPRESS_PREFIXLEN] = { .type = NLA_U32 }, \
[FRA_SUPPRESS_IFGROUP] = { .type = NLA_U32 }, \
[FRA_GOTO] = { .type = NLA_U32 }, \
- [FRA_L3MDEV] = { .type = NLA_U8 }
+ [FRA_L3MDEV] = { .type = NLA_U8 }, \
+ [FRA_UID_RANGE] = { .len = sizeof(struct fib_rule_uid_range) }
static inline void fib_rule_get(struct fib_rule *rule)
{
diff --git a/include/net/flow.h b/include/net/flow.h
index 035aa77..51373f3 100644
--- a/include/net/flow.h
+++ b/include/net/flow.h
@@ -11,6 +11,7 @@
#include <linux/in6.h>
#include <linux/atomic.h>
#include <net/flow_dissector.h>
+#include <linux/uidgid.h>
/*
* ifindex generation is per-net namespace, and loopback is
@@ -37,6 +38,7 @@ struct flowi_common {
#define FLOWI_FLAG_SKIP_NH_OIF 0x04
__u32 flowic_secid;
struct flowi_tunnel flowic_tun_key;
+ kuid_t flowic_uid;
};
union flowi_uli {
@@ -74,6 +76,7 @@ struct flowi4 {
#define flowi4_flags __fl_common.flowic_flags
#define flowi4_secid __fl_common.flowic_secid
#define flowi4_tun_key __fl_common.flowic_tun_key
+#define flowi4_uid __fl_common.flowic_uid
/* (saddr,daddr) must be grouped, same order as in IP header */
__be32 saddr;
@@ -131,6 +134,7 @@ struct flowi6 {
#define flowi6_flags __fl_common.flowic_flags
#define flowi6_secid __fl_common.flowic_secid
#define flowi6_tun_key __fl_common.flowic_tun_key
+#define flowi6_uid __fl_common.flowic_uid
struct in6_addr daddr;
struct in6_addr saddr;
/* Note: flowi6_tos is encoded in flowlabel, too. */
@@ -176,6 +180,7 @@ struct flowi {
#define flowi_flags u.__fl_common.flowic_flags
#define flowi_secid u.__fl_common.flowic_secid
#define flowi_tun_key u.__fl_common.flowic_tun_key
+#define flowi_uid u.__fl_common.flowic_uid
} __attribute__((__aligned__(BITS_PER_LONG/8)));
static inline struct flowi *flowi4_to_flowi(struct flowi4 *fl4)
diff --git a/include/uapi/linux/fib_rules.h b/include/uapi/linux/fib_rules.h
index 14404b3..bbf02a6 100644
--- a/include/uapi/linux/fib_rules.h
+++ b/include/uapi/linux/fib_rules.h
@@ -29,6 +29,11 @@ struct fib_rule_hdr {
__u32 flags;
};
+struct fib_rule_uid_range {
+ __u32 start;
+ __u32 end;
+};
+
enum {
FRA_UNSPEC,
FRA_DST, /* destination address */
@@ -51,6 +56,7 @@ enum {
FRA_OIFNAME,
FRA_PAD,
FRA_L3MDEV, /* iif or oif is l3mdev goto its table */
+ FRA_UID_RANGE, /* UID range */
__FRA_MAX
};
diff --git a/include/uapi/linux/rtnetlink.h b/include/uapi/linux/rtnetlink.h
index 5a78be5..e14377f 100644
--- a/include/uapi/linux/rtnetlink.h
+++ b/include/uapi/linux/rtnetlink.h
@@ -318,6 +318,7 @@ enum rtattr_type_t {
RTA_ENCAP,
RTA_EXPIRES,
RTA_PAD,
+ RTA_UID,
__RTA_MAX
};
diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index be4629c..5de436a 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -18,6 +18,11 @@
#include <net/fib_rules.h>
#include <net/ip_tunnels.h>
+static const struct fib_kuid_range fib_kuid_range_unset = {
+ KUIDT_INIT(0),
+ KUIDT_INIT(~0),
+};
+
int fib_default_rule_add(struct fib_rules_ops *ops,
u32 pref, u32 table, u32 flags)
{
@@ -33,6 +38,7 @@ int fib_default_rule_add(struct fib_rules_ops *ops,
r->table = table;
r->flags = flags;
r->fr_net = ops->fro_net;
+ r->uid_range = fib_kuid_range_unset;
r->suppress_prefixlen = -1;
r->suppress_ifgroup = -1;
@@ -172,6 +178,34 @@ void fib_rules_unregister(struct fib_rules_ops *ops)
}
EXPORT_SYMBOL_GPL(fib_rules_unregister);
+static int uid_range_set(struct fib_kuid_range *range)
+{
+ return uid_valid(range->start) && uid_valid(range->end);
+}
+
+static struct fib_kuid_range nla_get_kuid_range(struct nlattr **tb)
+{
+ struct fib_rule_uid_range *in;
+ struct fib_kuid_range out;
+
+ in = (struct fib_rule_uid_range *)nla_data(tb[FRA_UID_RANGE]);
+
+ out.start = make_kuid(current_user_ns(), in->start);
+ out.end = make_kuid(current_user_ns(), in->end);
+
+ return out;
+}
+
+static int nla_put_uid_range(struct sk_buff *skb, struct fib_kuid_range *range)
+{
+ struct fib_rule_uid_range out = {
+ from_kuid_munged(current_user_ns(), range->start),
+ from_kuid_munged(current_user_ns(), range->end)
+ };
+
+ return nla_put(skb, FRA_UID_RANGE, sizeof(out), &out);
+}
+
static int fib_rule_match(struct fib_rule *rule, struct fib_rules_ops *ops,
struct flowi *fl, int flags,
struct fib_lookup_arg *arg)
@@ -193,6 +227,10 @@ static int fib_rule_match(struct fib_rule *rule, struct fib_rules_ops *ops,
if (rule->l3mdev && !l3mdev_fib_rule_match(rule->fr_net, fl, arg))
goto out;
+ if (uid_lt(fl->flowi_uid, rule->uid_range.start) ||
+ uid_gt(fl->flowi_uid, rule->uid_range.end))
+ goto out;
+
ret = ops->match(rule, fl, flags);
out:
return (rule->flags & FIB_RULE_INVERT) ? !ret : ret;
@@ -429,6 +467,21 @@ int fib_nl_newrule(struct sk_buff *skb, struct nlmsghdr *nlh)
if (rule->l3mdev && rule->table)
goto errout_free;
+ if (tb[FRA_UID_RANGE]) {
+ if (current_user_ns() != net->user_ns) {
+ err = -EPERM;
+ goto errout_free;
+ }
+
+ rule->uid_range = nla_get_kuid_range(tb);
+
+ if (!uid_range_set(&rule->uid_range) ||
+ !uid_lte(rule->uid_range.start, rule->uid_range.end))
+ goto errout_free;
+ } else {
+ rule->uid_range = fib_kuid_range_unset;
+ }
+
if ((nlh->nlmsg_flags & NLM_F_EXCL) &&
rule_exists(ops, frh, tb, rule)) {
err = -EEXIST;
@@ -497,6 +550,7 @@ int fib_nl_delrule(struct sk_buff *skb, struct nlmsghdr *nlh)
struct fib_rules_ops *ops = NULL;
struct fib_rule *rule, *tmp;
struct nlattr *tb[FRA_MAX+1];
+ struct fib_kuid_range range;
int err = -EINVAL;
if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*frh)))
@@ -516,6 +570,14 @@ int fib_nl_delrule(struct sk_buff *skb, struct nlmsghdr *nlh)
if (err < 0)
goto errout;
+ if (tb[FRA_UID_RANGE]) {
+ range = nla_get_kuid_range(tb);
+ if (!uid_range_set(&range))
+ goto errout;
+ } else {
+ range = fib_kuid_range_unset;
+ }
+
list_for_each_entry(rule, &ops->rules_list, list) {
if (frh->action && (frh->action != rule->action))
continue;
@@ -552,6 +614,11 @@ int fib_nl_delrule(struct sk_buff *skb, struct nlmsghdr *nlh)
(rule->l3mdev != nla_get_u8(tb[FRA_L3MDEV])))
continue;
+ if (uid_range_set(&range) &&
+ (!uid_eq(rule->uid_range.start, range.start) ||
+ !uid_eq(rule->uid_range.end, range.end)))
+ continue;
+
if (!ops->compare(rule, frh, tb))
continue;
@@ -619,7 +686,8 @@ static inline size_t fib_rule_nlmsg_size(struct fib_rules_ops *ops,
+ nla_total_size(4) /* FRA_SUPPRESS_IFGROUP */
+ nla_total_size(4) /* FRA_FWMARK */
+ nla_total_size(4) /* FRA_FWMASK */
- + nla_total_size_64bit(8); /* FRA_TUN_ID */
+ + nla_total_size_64bit(8) /* FRA_TUN_ID */
+ + nla_total_size(sizeof(struct fib_kuid_range));
if (ops->nlmsg_payload)
payload += ops->nlmsg_payload(rule);
@@ -679,7 +747,9 @@ static int fib_nl_fill_rule(struct sk_buff *skb, struct fib_rule *rule,
(rule->tun_id &&
nla_put_be64(skb, FRA_TUN_ID, rule->tun_id, FRA_PAD)) ||
(rule->l3mdev &&
- nla_put_u8(skb, FRA_L3MDEV, rule->l3mdev)))
+ nla_put_u8(skb, FRA_L3MDEV, rule->l3mdev)) ||
+ (uid_range_set(&rule->uid_range) &&
+ nla_put_uid_range(skb, &rule->uid_range)))
goto nla_put_failure;
if (rule->suppress_ifgroup != -1) {
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
index c3b8047..d93eea8 100644
--- a/net/ipv4/fib_frontend.c
+++ b/net/ipv4/fib_frontend.c
@@ -610,6 +610,7 @@ const struct nla_policy rtm_ipv4_policy[RTA_MAX + 1] = {
[RTA_FLOW] = { .type = NLA_U32 },
[RTA_ENCAP_TYPE] = { .type = NLA_U16 },
[RTA_ENCAP] = { .type = NLA_NESTED },
+ [RTA_UID] = { .type = NLA_U32 },
};
static int rtm_to_fib_config(struct net *net, struct sk_buff *skb,
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 4392db8..92e59a6 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2504,6 +2504,11 @@ static int rt_fill_info(struct net *net, __be32 dst, __be32 src, u32 table_id,
nla_put_u32(skb, RTA_MARK, fl4->flowi4_mark))
goto nla_put_failure;
+ if (!uid_eq(fl4->flowi4_uid, INVALID_UID) &&
+ nla_put_u32(skb, RTA_UID,
+ from_kuid_munged(current_user_ns(), fl4->flowi4_uid)))
+ goto nla_put_failure;
+
error = rt->dst.error;
if (rt_is_input_route(rt)) {
@@ -2556,6 +2561,7 @@ static int inet_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh)
int mark;
struct sk_buff *skb;
u32 table_id = RT_TABLE_MAIN;
+ kuid_t uid;
err = nlmsg_parse(nlh, sizeof(*rtm), tb, RTA_MAX, rtm_ipv4_policy);
if (err < 0)
@@ -2583,6 +2589,10 @@ static int inet_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh)
dst = tb[RTA_DST] ? nla_get_in_addr(tb[RTA_DST]) : 0;
iif = tb[RTA_IIF] ? nla_get_u32(tb[RTA_IIF]) : 0;
mark = tb[RTA_MARK] ? nla_get_u32(tb[RTA_MARK]) : 0;
+ if (tb[RTA_UID])
+ uid = make_kuid(current_user_ns(), nla_get_u32(tb[RTA_UID]));
+ else
+ uid = (iif ? INVALID_UID : current_uid());
memset(&fl4, 0, sizeof(fl4));
fl4.daddr = dst;
@@ -2590,6 +2600,7 @@ static int inet_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh)
fl4.flowi4_tos = rtm->rtm_tos;
fl4.flowi4_oif = tb[RTA_OIF] ? nla_get_u32(tb[RTA_OIF]) : 0;
fl4.flowi4_mark = mark;
+ fl4.flowi4_uid = uid;
if (iif) {
struct net_device *dev;
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 947ed1d..fdb9c87 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -2797,6 +2797,7 @@ static const struct nla_policy rtm_ipv6_policy[RTA_MAX+1] = {
[RTA_ENCAP_TYPE] = { .type = NLA_U16 },
[RTA_ENCAP] = { .type = NLA_NESTED },
[RTA_EXPIRES] = { .type = NLA_U32 },
+ [RTA_UID] = { .type = NLA_U32 },
};
static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh,
@@ -3371,6 +3372,12 @@ static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh)
if (tb[RTA_MARK])
fl6.flowi6_mark = nla_get_u32(tb[RTA_MARK]);
+ if (tb[RTA_UID])
+ fl6.flowi6_uid = make_kuid(current_user_ns(),
+ nla_get_u32(tb[RTA_UID]));
+ else
+ fl6.flowi6_uid = iif ? INVALID_UID : current_uid();
+
if (iif) {
struct net_device *dev;
int flags = 0;
--
2.8.0.rc3.226.g39d4020
^ permalink raw reply related
* [PATCH net-next v2 1/3] net: core: Add a UID field to struct sock.
From: Lorenzo Colitti @ 2016-11-03 17:23 UTC (permalink / raw)
To: netdev; +Cc: davem, ek, eric.dumazet, zenczykowski, Lorenzo Colitti
In-Reply-To: <1478193823-60516-1-git-send-email-lorenzo@google.com>
Protocol sockets (struct sock) don't have UIDs, but most of the
time, they map 1:1 to userspace sockets (struct socket) which do.
Various operations such as the iptables xt_owner match need
access to the "UID of a socket", and do so by following the
backpointer to the struct socket. This involves taking
sk_callback_lock and doesn't work when there is no socket
because userspace has already called close().
Simplify this by adding a sk_uid field to struct sock whose value
matches the UID of the corresponding struct socket. The semantics
are as follows:
1. Whenever sk_socket is non-null: sk_uid is the same as the UID
in sk_socket, i.e., matches the return value of sock_i_uid.
Specifically, the UID is set when userspace calls socket(),
fchown(), or accept().
2. When sk_socket is NULL, sk_uid is defined as follows:
- For a socket that no longer has a sk_socket because
userspace has called close(): the previous UID.
- For a cloned socket (e.g., an incoming connection that is
established but on which userspace has not yet called
accept): the UID of the socket it was cloned from.
- For a socket that has never had an sk_socket: UID 0 inside
the user namespace corresponding to the network namespace
the socket belongs to.
Kernel sockets created by sock_create_kern are a special case
of #1 and sk_uid is the user that created them. For kernel
sockets created at network namespace creation time, such as the
per-processor ICMP and TCP sockets, this is the user that created
the network namespace.
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
---
include/net/sock.h | 7 +++++++
net/core/sock.c | 5 ++++-
net/socket.c | 14 ++++++++++++++
3 files changed, 25 insertions(+), 1 deletion(-)
diff --git a/include/net/sock.h b/include/net/sock.h
index 93331a1..cf617ee 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -419,6 +419,7 @@ struct sock {
u32 sk_max_ack_backlog;
__u32 sk_priority;
__u32 sk_mark;
+ kuid_t sk_uid;
struct pid *sk_peer_pid;
const struct cred *sk_peer_cred;
long sk_rcvtimeo;
@@ -1664,6 +1665,7 @@ static inline void sock_graft(struct sock *sk, struct socket *parent)
sk->sk_wq = parent->wq;
parent->sk = sk;
sk_set_socket(sk, parent);
+ sk->sk_uid = SOCK_INODE(parent)->i_uid;
security_sock_graft(sk, parent);
write_unlock_bh(&sk->sk_callback_lock);
}
@@ -1671,6 +1673,11 @@ static inline void sock_graft(struct sock *sk, struct socket *parent)
kuid_t sock_i_uid(struct sock *sk);
unsigned long sock_i_ino(struct sock *sk);
+static inline kuid_t sock_net_uid(const struct net *net, const struct sock *sk)
+{
+ return sk ? sk->sk_uid : make_kuid(net->user_ns, 0);
+}
+
static inline u32 net_tx_rndhash(void)
{
u32 v = prandom_u32();
diff --git a/net/core/sock.c b/net/core/sock.c
index d8e4532e..40dbc13 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -2460,8 +2460,11 @@ void sock_init_data(struct socket *sock, struct sock *sk)
sk->sk_type = sock->type;
sk->sk_wq = sock->wq;
sock->sk = sk;
- } else
+ sk->sk_uid = SOCK_INODE(sock)->i_uid;
+ } else {
sk->sk_wq = NULL;
+ sk->sk_uid = make_kuid(sock_net(sk)->user_ns, 0);
+ }
rwlock_init(&sk->sk_callback_lock);
lockdep_set_class_and_name(&sk->sk_callback_lock,
diff --git a/net/socket.c b/net/socket.c
index 970a7ea..4ce33c3 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -518,8 +518,22 @@ static ssize_t sockfs_listxattr(struct dentry *dentry, char *buffer,
return used;
}
+int sockfs_setattr(struct dentry *dentry, struct iattr *iattr)
+{
+ int err = simple_setattr(dentry, iattr);
+
+ if (!err) {
+ struct socket *sock = SOCKET_I(d_inode(dentry));
+
+ sock->sk->sk_uid = iattr->ia_uid;
+ }
+
+ return err;
+}
+
static const struct inode_operations sockfs_inode_ops = {
.listxattr = sockfs_listxattr,
+ .setattr = sockfs_setattr,
};
/**
--
2.8.0.rc3.226.g39d4020
^ permalink raw reply related
* Re: [PATCH net-next v1 00/21] amd-xgbe: AMD XGBE driver updates 2016-11-01
From: Tom Lendacky @ 2016-11-03 17:23 UTC (permalink / raw)
To: David Miller; +Cc: netdev, f.fainelli
In-Reply-To: <20161103.131405.2110100650645460415.davem@davemloft.net>
On 11/03/2016 12:14 PM, David Miller wrote:
>
> 20+ patches is too many to submit at one time.
>
> Making huge patch series submissions puts an unreasonable burdon on
> those developers who might decide to review your work, including me.
>
> Please keep your series down to a small, reasonable, size. Perhaps
> 10-15 patches maximum.
Ok, I'll break this up into smaller submissions - new device prep
work and then the new device support.
Thanks,
Tom
>
> I will not be looking at nor applying this series.
>
^ permalink raw reply
* [PATCH net-next v2] net: inet: Support UID-based routing
From: Lorenzo Colitti @ 2016-11-03 17:23 UTC (permalink / raw)
To: netdev; +Cc: davem, ek, eric.dumazet, zenczykowski
This patchset adds support for per-UID routing. It allows the
administrator to configure rules such as:
ip rule add uidrange 100-200 lookup 123
This functionality has been in use by all Android devices since
5.0. It is primarily used to impose per-app routing policies (on
Android, every app has its own UID) without having to resort to
rerouting packets in iptables, which breaks getsockname() and
MTU/MSS calculation, and generally disrupts end-to-end
connectivity.
This patch series is similar to the code currently used on
Android, but has better correctness and performance because
it stores the UID in the socket instead of calling sock_i_uid.
This avoids contention on sk->sk_callback_lock, and makes it
possible to correctly route a socket on which userspace has
called close(), for which sock_i_uid will return 0.
Changes from v1:
- Don't set the UID in sk_clone_lock, it's already set by
sock_copy.
- For packets originated by kernel sockets, don't use the socket
UID. This is the UID that created the namespace, but it might
not be mapped in the namespace at all. Instead, use UID 0 in
the namespace, which is less surprising and consistent with
what happens in the root namespace.
- Fix UID routing of IPv4 and IPv6 SYN_RECV sockets.
- Fix UID routing of received IPv6 redirects.
^ permalink raw reply
* Re: net/netlink: null-ptr-deref in netlink_dump/lock_acquire
From: Andrey Konovalov @ 2016-11-03 17:21 UTC (permalink / raw)
To: Eric Dumazet
Cc: Herbert Xu, Andrew Morton, David Decotigny, David S. Miller,
Dmitry Ivanov, Eric Dumazet, Florian Westphal, Greg Rose,
Johannes Berg, Matti Vaittinen, Pravin B Shelar,
stephen hemminger, Tom Herbert, Tycho Andersen, LKML, netdev,
syzkaller, Kostya Serebryany, Alexander Potapenko
In-Reply-To: <CAAeHK+xUFKMfW4WoYgyjJeYJZm_gAg3Nr3S0+8M-5H+6a=RVKw@mail.gmail.com>
Hi,
Another report that looks related:
[ INFO: possible circular locking dependency detected ]
4.9.0-rc3+ #344 Not tainted
-------------------------------------------------------
syz-executor/25526 is trying to acquire lock:
([ 950.351060] &table[i].mutex
[<ffffffff82dcf798>] nfnl_lock+0x28/0x30 net/netfilter/nfnetlink.c:61
but task is already holding lock:
([ 950.351060] rtnl_mutex
[< inline >] rtnl_lock net/core/rtnetlink.c:70
[<ffffffff82c15efb>] rtnetlink_rcv+0x1b/0x40 net/core/rtnetlink.c:4032
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
:
[ 950.351060] [<ffffffff8120c52e>] lock_acquire+0x17e/0x340
kernel/locking/lockdep.c:3746
[ 950.351060] [< inline >] __mutex_lock_common
kernel/locking/mutex.c:521
[ 950.351060] [<ffffffff83fb7221>]
mutex_lock_nested+0xb1/0x890 kernel/locking/mutex.c:621
[ 950.351060] [<ffffffff82c11297>] rtnl_lock+0x17/0x20
net/core/rtnetlink.c:70
[ 950.351060] [<ffffffff839fe31f>]
nl80211_get_reg_dump+0x5f/0x400 net/wireless/nl80211.c:6171
[ 950.351060] [<ffffffff82dc18d8>] genl_lock_dumpit+0x68/0x90
net/netlink/genetlink.c:517
[ 950.351060] [<ffffffff82db6947>] netlink_dump+0x397/0xac0
net/netlink/af_netlink.c:2110
[ 950.351060] [<ffffffff82db99b1>]
__netlink_dump_start+0x501/0x770 net/netlink/af_netlink.c:2200
[ 950.351060] [<ffffffff82dc2cb8>]
genl_family_rcv_msg+0xad8/0xc80 net/netlink/genetlink.c:584
[ 950.351060] [<ffffffff82dc3016>] genl_rcv_msg+0x1b6/0x270
net/netlink/genetlink.c:658
[ 950.351060] [<ffffffff82dc10a0>] netlink_rcv_skb+0x2c0/0x3b0
net/netlink/af_netlink.c:2281
[ 950.351060] [<ffffffff82dc21c8>] genl_rcv+0x28/0x40
net/netlink/genetlink.c:669
[ 950.351060] [< inline >] netlink_unicast_kernel
net/netlink/af_netlink.c:1214
[ 950.351060] [<ffffffff82dbf959>] netlink_unicast+0x5a9/0x880
net/netlink/af_netlink.c:1240
[ 950.351060] [<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0
net/netlink/af_netlink.c:1786
[ 950.351060] [< inline >] sock_sendmsg_nosec net/socket.c:606
[ 950.351060] [<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110
net/socket.c:616
[ 950.351060] [<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0
net/socket.c:814
[ 950.351060] [< inline >] new_sync_write fs/read_write.c:499
[ 950.351060] [<ffffffff8151bd44>] __vfs_write+0x334/0x570
fs/read_write.c:512
[ 950.351060] [<ffffffff8151f85b>] vfs_write+0x17b/0x500
fs/read_write.c:560
[ 950.351060] [< inline >] SYSC_write fs/read_write.c:607
[ 950.351060] [<ffffffff81523184>] SyS_write+0xd4/0x1a0
fs/read_write.c:599
[ 950.351060] [<ffffffff83fc0281>] entry_SYSCALL_64_fastpath+0x1f/0xc2
:
[ 950.351060] [<ffffffff8120c52e>] lock_acquire+0x17e/0x340
kernel/locking/lockdep.c:3746
[ 950.351060] [< inline >] __mutex_lock_common
kernel/locking/mutex.c:521
[ 950.351060] [<ffffffff83fb7221>]
mutex_lock_nested+0xb1/0x890 kernel/locking/mutex.c:621
[ 950.351060] [< inline >] genl_lock net/netlink/genetlink.c:31
[ 950.351060] [<ffffffff82dc18b1>] genl_lock_dumpit+0x41/0x90
net/netlink/genetlink.c:516
[ 950.351060] [<ffffffff82db6947>] netlink_dump+0x397/0xac0
net/netlink/af_netlink.c:2110
[ 950.351060] [<ffffffff82db99b1>]
__netlink_dump_start+0x501/0x770 net/netlink/af_netlink.c:2200
[ 950.351060] [<ffffffff82dc2cb8>]
genl_family_rcv_msg+0xad8/0xc80 net/netlink/genetlink.c:584
[ 950.351060] [<ffffffff82dc3016>] genl_rcv_msg+0x1b6/0x270
net/netlink/genetlink.c:658
[ 950.351060] [<ffffffff82dc10a0>] netlink_rcv_skb+0x2c0/0x3b0
net/netlink/af_netlink.c:2281
[ 950.351060] [<ffffffff82dc21c8>] genl_rcv+0x28/0x40
net/netlink/genetlink.c:669
[ 950.351060] [< inline >] netlink_unicast_kernel
net/netlink/af_netlink.c:1214
[ 950.351060] [<ffffffff82dbf959>] netlink_unicast+0x5a9/0x880
net/netlink/af_netlink.c:1240
[ 950.351060] [<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0
net/netlink/af_netlink.c:1786
[ 950.351060] [< inline >] sock_sendmsg_nosec net/socket.c:606
[ 950.351060] [<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110
net/socket.c:616
[ 950.351060] [<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0
net/socket.c:814
[ 950.351060] [< inline >] new_sync_write fs/read_write.c:499
[ 950.351060] [<ffffffff8151bd44>] __vfs_write+0x334/0x570
fs/read_write.c:512
[ 950.351060] [<ffffffff8151f85b>] vfs_write+0x17b/0x500
fs/read_write.c:560
[ 950.351060] [< inline >] SYSC_write fs/read_write.c:607
[ 950.351060] [<ffffffff81523184>] SyS_write+0xd4/0x1a0
fs/read_write.c:599
[ 950.351060] [<ffffffff83fc0281>] entry_SYSCALL_64_fastpath+0x1f/0xc2
:
[ 950.351060] [<ffffffff8120c52e>] lock_acquire+0x17e/0x340
kernel/locking/lockdep.c:3746
[ 950.351060] [< inline >] __mutex_lock_common
kernel/locking/mutex.c:521
[ 950.351060] [<ffffffff83fb7221>]
mutex_lock_nested+0xb1/0x890 kernel/locking/mutex.c:621
[ 950.351060] [<ffffffff82db95aa>]
__netlink_dump_start+0xfa/0x770 net/netlink/af_netlink.c:2170
[ 950.351060] [< inline >] netlink_dump_start
include/linux/netlink.h:165
[ 950.351060] [<ffffffff82e19309>]
ctnetlink_stat_ct_cpu+0xd9/0x130
net/netfilter/nf_conntrack_netlink.c:2045
[ 950.351060] [<ffffffff82dd1d80>]
nfnetlink_rcv_msg+0xa10/0xc10 net/netfilter/nfnetlink.c:212
[ 950.351060] [<ffffffff82dc10a0>] netlink_rcv_skb+0x2c0/0x3b0
net/netlink/af_netlink.c:2281
[ 950.351060] [<ffffffff82dd07e8>] nfnetlink_rcv+0x848/0x1170
net/netfilter/nfnetlink.c:474
[ 950.351060] [< inline >] netlink_unicast_kernel
net/netlink/af_netlink.c:1214
[ 950.351060] [<ffffffff82dbf959>] netlink_unicast+0x5a9/0x880
net/netlink/af_netlink.c:1240
[ 950.351060] [<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0
net/netlink/af_netlink.c:1786
[ 950.351060] [< inline >] sock_sendmsg_nosec net/socket.c:606
[ 950.351060] [<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110
net/socket.c:616
[ 950.351060] [<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0
net/socket.c:814
[ 950.351060] [< inline >] new_sync_write fs/read_write.c:499
[ 950.351060] [<ffffffff8151bd44>] __vfs_write+0x334/0x570
fs/read_write.c:512
[ 950.351060] [<ffffffff8151f85b>] vfs_write+0x17b/0x500
fs/read_write.c:560
[ 950.351060] [< inline >] SYSC_write fs/read_write.c:607
[ 950.351060] [<ffffffff81523184>] SyS_write+0xd4/0x1a0
fs/read_write.c:599
[ 950.351060] [<ffffffff83fc0281>] entry_SYSCALL_64_fastpath+0x1f/0xc2
:
[ 950.351060] [< inline >] check_prev_add
kernel/locking/lockdep.c:1829
[ 950.351060] [< inline >] check_prevs_add
kernel/locking/lockdep.c:1939
[ 950.351060] [< inline >] validate_chain
kernel/locking/lockdep.c:2266
[ 950.351060] [<ffffffff8120b0ca>]
__lock_acquire+0x2b4a/0x3450 kernel/locking/lockdep.c:3335
[ 950.351060] [<ffffffff8120c52e>] lock_acquire+0x17e/0x340
kernel/locking/lockdep.c:3746
[ 950.351060] [< inline >] __mutex_lock_common
kernel/locking/mutex.c:521
[ 950.351060] [<ffffffff83fb7221>]
mutex_lock_nested+0xb1/0x890 kernel/locking/mutex.c:621
[ 950.351060] [<ffffffff82dcf798>] nfnl_lock+0x28/0x30
net/netfilter/nfnetlink.c:61
[ 950.351060] [<ffffffff82e847bf>]
nf_tables_netdev_event+0x14f/0x590
net/netfilter/nf_tables_netdev.c:122
[ 950.351060] [<ffffffff8117c131>]
notifier_call_chain+0x91/0x1a0 kernel/notifier.c:93
[ 950.351060] [< inline >] __raw_notifier_call_chain
kernel/notifier.c:394
[ 950.351060] [<ffffffff8117c2bd>]
raw_notifier_call_chain+0x2d/0x40 kernel/notifier.c:401
[ 950.351060] [<ffffffff82bc3851>]
call_netdevice_notifiers_info+0x51/0x90 net/core/dev.c:1645
[ 950.351060] [< inline >] call_netdevice_notifiers
net/core/dev.c:1661
[ 950.351060] [<ffffffff82be77c7>] dev_change_name+0x5d7/0x920
net/core/dev.c:1204
[ 950.351060] [<ffffffff82c171dd>] do_setlink+0x83d/0x30d0
net/core/rtnetlink.c:1993
[ 950.351060] [<ffffffff82c19ce0>] rtnl_setlink+0x270/0x330
net/core/rtnetlink.c:2231
[ 950.351060] [<ffffffff82c254a4>]
rtnetlink_rcv_msg+0x274/0x700 net/core/rtnetlink.c:4027
[ 950.351060] [<ffffffff82dc10a0>] netlink_rcv_skb+0x2c0/0x3b0
net/netlink/af_netlink.c:2281
[ 950.351060] [<ffffffff82c15f0a>] rtnetlink_rcv+0x2a/0x40
net/core/rtnetlink.c:4033
[ 950.351060] [< inline >] netlink_unicast_kernel
net/netlink/af_netlink.c:1214
[ 950.351060] [<ffffffff82dbf959>] netlink_unicast+0x5a9/0x880
net/netlink/af_netlink.c:1240
[ 950.351060] [<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0
net/netlink/af_netlink.c:1786
[ 950.351060] [< inline >] sock_sendmsg_nosec net/socket.c:606
[ 950.351060] [<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110
net/socket.c:616
[ 950.351060] [<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0
net/socket.c:814
[ 950.351060] [< inline >] new_sync_write fs/read_write.c:499
[ 950.351060] [<ffffffff8151bd44>] __vfs_write+0x334/0x570
fs/read_write.c:512
[ 950.351060] [<ffffffff8151f85b>] vfs_write+0x17b/0x500
fs/read_write.c:560
[ 950.351060] [< inline >] SYSC_write fs/read_write.c:607
[ 950.351060] [<ffffffff81523184>] SyS_write+0xd4/0x1a0
fs/read_write.c:599
[ 950.351060] [<ffffffff81006465>] do_syscall_64+0x195/0x490
arch/x86/entry/common.c:280
[ 950.351060] [<ffffffff83fc0349>] return_from_SYSCALL_64+0x0/0x7a
other info that might help us debug this:
Chain exists of:
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock([ 950.351060] rtnl_mutex
);
lock([ 950.351060] genl_mutex
);
lock([ 950.351060] rtnl_mutex
);
lock([ 950.351060] &table[i].mutex
);
*** DEADLOCK ***
1 lock held by syz-executor/25526:
#0: [ 950.351060] (
stack backtrace:
CPU: 0 PID: 25526 Comm: syz-executor Not tainted 4.9.0-rc3+ #344
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
ffff8800692b6d30 ffffffff81b46934 ffffffff859bdf80 ffffffff859de740
ffffffff859b3870 1ffffffff0bade82 ffff8800692b6d80 ffffffff81204cfd
ffffffff859b3870 ffff88006859cca0 000000006859c480 ffff88006859cca0
Call Trace:
[< inline >] __dump_stack lib/dump_stack.c:15
[<ffffffff81b46934>] dump_stack+0xb3/0x10f lib/dump_stack.c:51
[<ffffffff81204cfd>] print_circular_bug+0x28d/0x350
kernel/locking/lockdep.c:1202
[< inline >] check_prev_add kernel/locking/lockdep.c:1829
[< inline >] check_prevs_add kernel/locking/lockdep.c:1939
[< inline >] validate_chain kernel/locking/lockdep.c:2266
[<ffffffff8120b0ca>] __lock_acquire+0x2b4a/0x3450 kernel/locking/lockdep.c:3335
[<ffffffff8120c52e>] lock_acquire+0x17e/0x340 kernel/locking/lockdep.c:3746
[< inline >] __mutex_lock_common kernel/locking/mutex.c:521
[<ffffffff83fb7221>] mutex_lock_nested+0xb1/0x890 kernel/locking/mutex.c:621
[<ffffffff82dcf798>] nfnl_lock+0x28/0x30 net/netfilter/nfnetlink.c:61
[<ffffffff82e847bf>] nf_tables_netdev_event+0x14f/0x590
net/netfilter/nf_tables_netdev.c:122
[<ffffffff8117c131>] notifier_call_chain+0x91/0x1a0 kernel/notifier.c:93
[< inline >] __raw_notifier_call_chain kernel/notifier.c:394
[<ffffffff8117c2bd>] raw_notifier_call_chain+0x2d/0x40 kernel/notifier.c:401
[<ffffffff82bc3851>] call_netdevice_notifiers_info+0x51/0x90
net/core/dev.c:1645
[< inline >] call_netdevice_notifiers net/core/dev.c:1661
[<ffffffff82be77c7>] dev_change_name+0x5d7/0x920 net/core/dev.c:1204
[<ffffffff82c171dd>] do_setlink+0x83d/0x30d0 net/core/rtnetlink.c:1993
[<ffffffff82c19ce0>] rtnl_setlink+0x270/0x330 net/core/rtnetlink.c:2231
[<ffffffff82c254a4>] rtnetlink_rcv_msg+0x274/0x700 net/core/rtnetlink.c:4027
[<ffffffff82dc10a0>] netlink_rcv_skb+0x2c0/0x3b0 net/netlink/af_netlink.c:2281
[<ffffffff82c15f0a>] rtnetlink_rcv+0x2a/0x40 net/core/rtnetlink.c:4033
[< inline >] netlink_unicast_kernel net/netlink/af_netlink.c:1214
[<ffffffff82dbf959>] netlink_unicast+0x5a9/0x880 net/netlink/af_netlink.c:1240
[<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0 net/netlink/af_netlink.c:1786
[< inline >] sock_sendmsg_nosec net/socket.c:606
[<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110 net/socket.c:616
[<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0 net/socket.c:814
[< inline >] new_sync_write fs/read_write.c:499
[<ffffffff8151bd44>] __vfs_write+0x334/0x570 fs/read_write.c:512
[<ffffffff8151f85b>] vfs_write+0x17b/0x500 fs/read_write.c:560
[< inline >] SYSC_write fs/read_write.c:607
[<ffffffff81523184>] SyS_write+0xd4/0x1a0 fs/read_write.c:599
[<ffffffff81006465>] do_syscall_64+0x195/0x490 arch/x86/entry/common.c:280
[<ffffffff83fc0349>] entry_SYSCALL64_slow_path+0x25/0x25
On Thu, Nov 3, 2016 at 4:08 AM, Andrey Konovalov <andreyknvl@google.com> wrote:
> Hi Eric,
>
> This fixes the second report, the first one is still there.
> Apparently these are two separate issues.
>
> For the second one:
> Tested-by: Andrey Konovalov <andreyknvl@google.com>
>
> Thanks for the fix!
>
> On Thu, Nov 3, 2016 at 3:58 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
>> On Thu, 2016-11-03 at 03:36 +0100, Andrey Konovalov wrote:
>>> On Thu, Nov 3, 2016 at 1:15 AM, Andrey Konovalov <andreyknvl@google.com> wrote:
>>> > On Wed, Oct 19, 2016 at 4:13 PM, Andrey Konovalov <andreyknvl@google.com> wrote:
>>> >> Hi,
>>> >>
>>> >> I've got the following error report while running the syzkaller fuzzer:
>>> >>
>>> >> kasan: CONFIG_KASAN_INLINE enabled
>>> >> kasan: GPF could be caused by NULL-ptr deref or user memory access
>>> >> general protection fault: 0000 [#1] SMP KASAN
>>> >> Modules linked in:
>>> >> CPU: 1 PID: 3933 Comm: syz-executor Not tainted 4.9.0-rc1+ #230
>>> >> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>>> >> task: ffff88006b79d800 task.stack: ffff88006bbc0000
>>> >> RIP: 0010:[<ffffffff8120872d>] [<ffffffff8120872d>]
>>> >> __lock_acquire+0x12d/0x3450 kernel/locking/lockdep.c:3221
>>> >> RSP: 0018:ffff88006bbc7420 EFLAGS: 00010006
>>> >> RAX: 0000000000000046 RBX: dffffc0000000000 RCX: 0000000000000000
>>> >> RDX: 000000000000000c RSI: 0000000000000000 RDI: 0000000000000003
>>> >> RBP: ffff88006bbc75c0 R08: 0000000000000001 R09: 0000000000000000
>>> >> R10: 0000000000000000 R11: ffffffff85f42240 R12: ffff88006b79d800
>>> >> R13: ffffffff84bfe4e0 R14: 0000000000000001 R15: 0000000000000060
>>> >> FS: 00007fd9c41cc700(0000) GS:ffff88006cd00000(0000) knlGS:0000000000000000
>>> >> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>>> >> CR2: 0000000000451f80 CR3: 00000000638f0000 CR4: 00000000000006e0
>>> >> Stack:
>>> >> 0000000000000000 ffff88006bbc0000 ffff88006bbc8000 0000000000000000
>>> >> 0000000000000002 ffff88006b79d800 0000000000000000 ffff88006bbc7f48
>>> >> ffffffff852adc60 0000000000000000 ffffffff852adc64 1ffffffff0b40135
>>> >> Call Trace:
>>> >> [<ffffffff8120c5ae>] lock_acquire+0x17e/0x340 kernel/locking/lockdep.c:3746
>>> >> [< inline >] __mutex_lock_common kernel/locking/mutex.c:521
>>> >> [<ffffffff83fb6fe1>] mutex_lock_nested+0xb1/0x890 kernel/locking/mutex.c:621
>>> >> [<ffffffff82db6fd0>] netlink_dump+0x50/0xac0 net/netlink/af_netlink.c:2067
>>> >> [<ffffffff82dba381>] __netlink_dump_start+0x501/0x770
>>> >> net/netlink/af_netlink.c:2200
>>> >> [<ffffffff82dc35b2>] genl_family_rcv_msg+0xa02/0xc80
>>> >> net/netlink/genetlink.c:595
>>> >> [<ffffffff82dc39e6>] genl_rcv_msg+0x1b6/0x270 net/netlink/genetlink.c:658
>>> >> [<ffffffff82dc1a70>] netlink_rcv_skb+0x2c0/0x3b0 net/netlink/af_netlink.c:2281
>>> >> [<ffffffff82dc2b98>] genl_rcv+0x28/0x40 net/netlink/genetlink.c:669
>>> >> [< inline >] netlink_unicast_kernel net/netlink/af_netlink.c:1214
>>> >> [<ffffffff82dc0329>] netlink_unicast+0x5a9/0x880 net/netlink/af_netlink.c:1240
>>> >> [<ffffffff82dc0fb7>] netlink_sendmsg+0x9b7/0xce0 net/netlink/af_netlink.c:1786
>>> >> [< inline >] sock_sendmsg_nosec net/socket.c:606
>>> >> [<ffffffff82b7075c>] sock_sendmsg+0xcc/0x110 net/socket.c:616
>>> >> [<ffffffff82b709c1>] sock_write_iter+0x221/0x3b0 net/socket.c:814
>>> >> [< inline >] new_sync_write fs/read_write.c:499
>>> >> [<ffffffff8151c944>] __vfs_write+0x334/0x570 fs/read_write.c:512
>>> >> [<ffffffff8152045b>] vfs_write+0x17b/0x500 fs/read_write.c:560
>>> >> [< inline >] SYSC_write fs/read_write.c:607
>>> >> [<ffffffff81523d84>] SyS_write+0xd4/0x1a0 fs/read_write.c:599
>>> >> [<ffffffff83fc0141>] entry_SYSCALL_64_fastpath+0x1f/0xc2
>>> >> arch/x86/entry/entry_64.S:209
>>> >> Code: 0f 1f 44 00 00 f6 c4 02 0f 85 24 0a 00 00 44 8b 35 c9 61 8b 03
>>> >> 45 85 f6 74 2c 4c 89 fa 48 bb 00 00 00 00 00 fc ff df 48 c1 ea 03 <80>
>>> >> 3c 1a 00 0f 85 04 2f 00 00 49 81 3f a0 dc 2a 85 41 be 00 00
>>> >> RIP [<ffffffff8120872d>] __lock_acquire+0x12d/0x3450
>>> >> kernel/locking/lockdep.c:3221
>>> >> RSP <ffff88006bbc7420>
>>> >> ---[ end trace 685b3c182bf7f25c ]---
>>> >>
>>> >> The reproducer is attached.
>>> >>
>>> >> On commit 1a1891d762d6e64daf07b5be4817e3fbb29e3c59 (Oct 18).
>>> >
>>> > (Adding more maintainers)
>>> >
>>> > Still seeing this on 0c183d92b20b5c84ca655b45ef57b3318b83eb9e (Oct 31).
>>>
>>> Here is another report that might be related:
>>>
>>> =====================================
>>> [ BUG: bad unlock balance detected! ]
>>> 4.9.0-rc3+ #336 Not tainted
>>> -------------------------------------
>>> syz-executor/4018 is trying to release lock ([ 36.220068] nl_table_lock
>>> ) at:
>>> [<ffffffff82dc8683>] netlink_diag_dump+0x1a3/0x250 net/netlink/diag.c:182
>>> but there are no more locks to release!
>>>
>>> other info that might help us debug this:
>>> 3 locks held by syz-executor/4018:
>>> #0: [ 36.220068] (
>>> sock_diag_mutex[ 36.220068] ){+.+.+.}
>>> , at: [ 36.220068] [<ffffffff82c3873b>] sock_diag_rcv+0x1b/0x40
>>> #1: [ 36.220068] (
>>> sock_diag_table_mutex[ 36.220068] ){+.+.+.}
>>> , at: [ 36.220068] [<ffffffff82c38e00>] sock_diag_rcv_msg+0x140/0x3a0
>>> #2: [ 36.220068] (
>>> nlk->cb_mutex[ 36.220068] ){+.+.+.}
>>> , at: [ 36.220068] [<ffffffff82db6600>] netlink_dump+0x50/0xac0
>>>
>>> stack backtrace:
>>> CPU: 1 PID: 4018 Comm: syz-executor Not tainted 4.9.0-rc3+ #336
>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>>> ffff8800645df688 ffffffff81b46934 ffffffff84eb3e78 ffff88006ad85800
>>> ffffffff82dc8683 ffffffff84eb3e78 ffff8800645df6b8 ffffffff812043ca
>>> dffffc0000000000 ffff88006ad85ff8 ffff88006ad85fd0 00000000ffffffff
>>> Call Trace:
>>> [< inline >] __dump_stack lib/dump_stack.c:15
>>> [<ffffffff81b46934>] dump_stack+0xb3/0x10f lib/dump_stack.c:51
>>> [<ffffffff812043ca>] print_unlock_imbalance_bug+0x17a/0x1a0
>>> kernel/locking/lockdep.c:3388
>>> [< inline >] __lock_release kernel/locking/lockdep.c:3512
>>> [<ffffffff8120cfd8>] lock_release+0x8e8/0xc60 kernel/locking/lockdep.c:3765
>>> [< inline >] __raw_read_unlock ./include/linux/rwlock_api_smp.h:225
>>> [<ffffffff83fc001a>] _raw_read_unlock+0x1a/0x30 kernel/locking/spinlock.c:255
>>> [<ffffffff82dc8683>] netlink_diag_dump+0x1a3/0x250 net/netlink/diag.c:182
>>> [<ffffffff82db6947>] netlink_dump+0x397/0xac0 net/netlink/af_netlink.c:2110
>>> [<ffffffff82db99b1>] __netlink_dump_start+0x501/0x770
>>> net/netlink/af_netlink.c:2200
>>> [< inline >] netlink_dump_start ./include/linux/netlink.h:165
>>> [<ffffffff82dc75d1>] netlink_diag_handler_dump+0x191/0x220
>>> net/netlink/diag.c:218
>>> [< inline >] __sock_diag_cmd net/core/sock_diag.c:239
>>> [<ffffffff82c38fd6>] sock_diag_rcv_msg+0x316/0x3a0 net/core/sock_diag.c:270
>>> [<ffffffff82dc10a0>] netlink_rcv_skb+0x2c0/0x3b0 net/netlink/af_netlink.c:2281
>>> [<ffffffff82c3874a>] sock_diag_rcv+0x2a/0x40 net/core/sock_diag.c:281
>>> [< inline >] netlink_unicast_kernel net/netlink/af_netlink.c:1214
>>> [<ffffffff82dbf959>] netlink_unicast+0x5a9/0x880 net/netlink/af_netlink.c:1240
>>> [<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0 net/netlink/af_netlink.c:1786
>>> [< inline >] sock_sendmsg_nosec net/socket.c:606
>>> [<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110 net/socket.c:616
>>> [<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0 net/socket.c:814
>>> [< inline >] new_sync_write fs/read_write.c:499
>>> [<ffffffff8151bd44>] __vfs_write+0x334/0x570 fs/read_write.c:512
>>> [<ffffffff8151f85b>] vfs_write+0x17b/0x500 fs/read_write.c:560
>>> [< inline >] SYSC_write fs/read_write.c:607
>>> [<ffffffff81523184>] SyS_write+0xd4/0x1a0 fs/read_write.c:599
>>> [<ffffffff83fc0401>] entry_SYSCALL_64_fastpath+0x1f/0xc2
>>> arch/x86/entry/entry_64.S:209
>>> ------------[ cut here ]------------
>>> WARNING: CPU: 1 PID: 4018 at net/core/skbuff.c:654[< none
>>> >] skb_release_head_state+0x1ca/0x240 net/core/skbuff.c:654
>>> Modules linked in:
>>> CPU: 1 PID: 4018 Comm: syz-executor Not tainted 4.9.0-rc3+ #336
>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>>> ffff8800645df920 ffffffff81b46934 0000000000000000 0000000000000000
>>> ffffffff84401fa0 0000000000000000 ffff8800645df968 ffffffff811112f7
>>> ffffffff83fb92f2 ffff88000000028e ffffffff84401fa0 000000000000028e
>>> Call Trace:
>>> [< inline >] __dump_stack lib/dump_stack.c:15
>>> [<ffffffff81b46934>] dump_stack+0xb3/0x10f lib/dump_stack.c:51
>>> [<ffffffff811112f7>] __warn+0x1a7/0x1f0 kernel/panic.c:550
>>> [<ffffffff8111150c>] warn_slowpath_null+0x2c/0x40 kernel/panic.c:585
>>> [<ffffffff82b885ea>] skb_release_head_state+0x1ca/0x240 net/core/skbuff.c:654
>>> [<ffffffff82b91815>] skb_release_all+0x15/0x60 net/core/skbuff.c:668
>>> [< inline >] __kfree_skb net/core/skbuff.c:684
>>> [<ffffffff82ba0175>] consume_skb+0x115/0x2e0 net/core/skbuff.c:757
>>> [< inline >] netlink_unicast_kernel net/netlink/af_netlink.c:1215
>>> [<ffffffff82dbf961>] netlink_unicast+0x5b1/0x880 net/netlink/af_netlink.c:1240
>>> [<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0 net/netlink/af_netlink.c:1786
>>> [< inline >] sock_sendmsg_nosec net/socket.c:606
>>> [<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110 net/socket.c:616
>>> [<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0 net/socket.c:814
>>> [< inline >] new_sync_write fs/read_write.c:499
>>> [<ffffffff8151bd44>] __vfs_write+0x334/0x570 fs/read_write.c:512
>>> [<ffffffff8151f85b>] vfs_write+0x17b/0x500 fs/read_write.c:560
>>> [< inline >] SYSC_write fs/read_write.c:607
>>> [<ffffffff81523184>] SyS_write+0xd4/0x1a0 fs/read_write.c:599
>>> [<ffffffff83fc0401>] entry_SYSCALL_64_fastpath+0x1f/0xc2
>>> arch/x86/entry/entry_64.S:209
>>> ---[ end trace bb9fa7cf182d59a5 ]---
>>> BUG: scheduling while atomic: syz-executor/4018/0x7fffffff
>>> INFO: lockdep is turned off.
>>> Modules linked in:
>>> CPU: 1 PID: 4018 Comm: syz-executor Tainted: G W 4.9.0-rc3+ #336
>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>>> ffff8800645dfe28 ffffffff81b46934 dffffc0000000000 000000007fffffff
>>> 00000000000214c0 0000000000000001 ffff8800645dfe48 ffffffff8119113a
>>> ffff88006cd214c0 0000000000000000 ffff8800645dfec8 ffffffff83fb030a
>>> Call Trace:
>>> [< inline >] __dump_stack lib/dump_stack.c:15
>>> [<ffffffff81b46934>] dump_stack+0xb3/0x10f lib/dump_stack.c:51
>>> [<ffffffff8119113a>] __schedule_bug+0xfa/0x140 kernel/sched/core.c:3230
>>> [< inline >] schedule_debug kernel/sched/core.c:3245
>>> [<ffffffff83fb030a>] __schedule+0xfda/0x1ab0 kernel/sched/core.c:3345
>>> [<ffffffff83fb0e70>] schedule+0x90/0x1b0 kernel/sched/core.c:3457
>>> [<ffffffff810039e9>] exit_to_usermode_loop+0xc9/0x130
>>> arch/x86/entry/common.c:149
>>> [< inline >] prepare_exit_to_usermode arch/x86/entry/common.c:190
>>> [<ffffffff81006298>] syscall_return_slowpath+0x1a8/0x1e0
>>> arch/x86/entry/common.c:259
>>> [<ffffffff83fc04a2>] entry_SYSCALL_64_fastpath+0xc0/0xc2
>>> arch/x86/entry/entry_64.S:244
>>> NOHZ: local_softirq_pending 202
>>> ------------[ cut here ]------------
>>> WARNING: CPU: 1 PID: 4018 at net/core/skbuff.c:654[< none
>>> >] skb_release_head_state+0x1ca/0x240 net/core/skbuff.c:654
>>> Modules linked in:[ 36.328353] CPU: 1 PID: 4018 Comm: syz-executor
>>> Tainted: G W 4.9.0-rc3+ #336
>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>>> ffff8800645df920 ffffffff81b46934 0000000000000000 0000000000000000
>>> ffffffff84401fa0 0000000000000000 ffff8800645df968 ffffffff811112f7
>>> ffffffff83fb92f2 ffff88000000028e ffffffff84401fa0 000000000000028e
>>> Call Trace:
>>> [< inline >] __dump_stack lib/dump_stack.c:15
>>> [<ffffffff81b46934>] dump_stack+0xb3/0x10f lib/dump_stack.c:51
>>> [<ffffffff811112f7>] __warn+0x1a7/0x1f0 kernel/panic.c:550
>>> [<ffffffff8111150c>] warn_slowpath_null+0x2c/0x40 kernel/panic.c:585
>>> [<ffffffff82b885ea>] skb_release_head_state+0x1ca/0x240 net/core/skbuff.c:654
>>> [<ffffffff82b91815>] skb_release_all+0x15/0x60 net/core/skbuff.c:668
>>> [< inline >] __kfree_skb net/core/skbuff.c:684
>>> [<ffffffff82ba0175>] consume_skb+0x115/0x2e0 net/core/skbuff.c:757
>>> [< inline >] netlink_unicast_kernel net/netlink/af_netlink.c:1215
>>> [<ffffffff82dbf961>] netlink_unicast+0x5b1/0x880 net/netlink/af_netlink.c:1240
>>> [<ffffffff82dc05e7>] netlink_sendmsg+0x9b7/0xce0 net/netlink/af_netlink.c:1786
>>> [< inline >] sock_sendmsg_nosec net/socket.c:606
>>> [<ffffffff82b6f75c>] sock_sendmsg+0xcc/0x110 net/socket.c:616
>>> [<ffffffff82b6f9c1>] sock_write_iter+0x221/0x3b0 net/socket.c:814
>>> [< inline >] new_sync_write fs/read_write.c:499
>>> [<ffffffff8151bd44>] __vfs_write+0x334/0x570 fs/read_write.c:512
>>> [<ffffffff8151f85b>] vfs_write+0x17b/0x500 fs/read_write.c:560
>>> [< inline >] SYSC_write fs/read_write.c:607
>>> [<ffffffff81523184>] SyS_write+0xd4/0x1a0 fs/read_write.c:599
>>> [<ffffffff83fc0401>] entry_SYSCALL_64_fastpath+0x1f/0xc2
>>> arch/x86/entry/entry_64.S:209
>>> ---[ end trace bb9fa7cf182d59a6 ]---
>>> BUG: sleeping function called from invalid context at
>>> ./include/linux/freezer.h:56
>>> in_atomic(): 1, irqs_disabled(): 0, pid: 4018, name: syz-executor
>>> INFO: lockdep is turned off.
>>> CPU: 1 PID: 4018 Comm: syz-executor Tainted: G W 4.9.0-rc3+ #336
>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>>> ffff8800645dfbb0 ffffffff81b46934 ffff88006ad85800 ffff8800645d8000
>>> ffff88006ad85800 0000000000000000 ffff8800645dfbd8 ffffffff81192131
>>> ffff88006ad85800 ffffffff8404c140 0000000000000038 ffff8800645dfc18
>>> Call Trace:
>>> [< inline >] __dump_stack lib/dump_stack.c:15
>>> [<ffffffff81b46934>] dump_stack+0xb3/0x10f lib/dump_stack.c:51
>>> [<ffffffff81192131>] ___might_sleep+0x281/0x3c0 kernel/sched/core.c:7767
>>> [<ffffffff81192306>] __might_sleep+0x96/0x1a0 kernel/sched/core.c:7726
>>> [< inline >] try_to_freeze_unsafe ./include/linux/freezer.h:56
>>> [< inline >] try_to_freeze ./include/linux/freezer.h:66
>>> [<ffffffff81143849>] get_signal+0x129/0x15a0 kernel/signal.c:2147
>>> [<ffffffff81054aad>] do_signal+0x8d/0x1a30 arch/x86/kernel/signal.c:807
>>> [<ffffffff81003a05>] exit_to_usermode_loop+0xe5/0x130
>>> arch/x86/entry/common.c:156
>>> [< inline >] prepare_exit_to_usermode arch/x86/entry/common.c:190
>>> [<ffffffff81006298>] syscall_return_slowpath+0x1a8/0x1e0
>>> arch/x86/entry/common.c:259
>>> [<ffffffff83fc04a2>] entry_SYSCALL_64_fastpath+0xc0/0xc2
>>> arch/x86/entry/entry_64.S:244
>>> Kernel panic - not syncing: Aiee, killing interrupt handler!
>>> CPU: 1 PID: 4018 Comm: syz-executor Tainted: G W 4.9.0-rc3+ #336
>>> Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
>>> ffff8800645df998 ffffffff81b46934 0000000000000003 dffffc0000000000
>>> dffffc0000000000 ffff8800645dfa04 ffff8800645dfa60 ffffffff8140bf7a
>>> 0000000041b58ab3 ffffffff84797a7d ffffffff8140bdbe ffffffff00000000
>>> Call Trace:
>>> [< inline >] __dump_stack lib/dump_stack.c:15
>>> [<ffffffff81b46934>] dump_stack+0xb3/0x10f lib/dump_stack.c:51
>>> [<ffffffff8140bf7a>] panic+0x1bc/0x39d kernel/panic.c:179
>>> [<ffffffff8111cfd8>] do_exit+0x1b48/0x2ac0 kernel/exit.c:740
>>> [<ffffffff811222be>] do_group_exit+0x10e/0x340 kernel/exit.c:931
>>> [<ffffffff81143d54>] get_signal+0x634/0x15a0 kernel/signal.c:2307
>>> [<ffffffff81054aad>] do_signal+0x8d/0x1a30 arch/x86/kernel/signal.c:807
>>> [<ffffffff81003a05>] exit_to_usermode_loop+0xe5/0x130
>>> arch/x86/entry/common.c:156
>>> [< inline >] prepare_exit_to_usermode arch/x86/entry/common.c:190
>>> [<ffffffff81006298>] syscall_return_slowpath+0x1a8/0x1e0
>>> arch/x86/entry/common.c:259
>>> [<ffffffff83fc04a2>] entry_SYSCALL_64_fastpath+0xc0/0xc2
>>> arch/x86/entry/entry_64.S:244
>>> Kernel Offset: disabled
>>> ---[ end Kernel panic - not syncing: Aiee, killing interrupt handler!
>>
>>
>> This is probably a leftover after commit
>> ad202074320cd75b31b8cdb58cca0d4ef6aaea8a
>> ("netlink: Use rhashtable walk interface in diag dump")
>>
>> Please try this trivial fix :
>>
>> diff --git a/net/netlink/diag.c b/net/netlink/diag.c
>> index b2f0e986a6f4..a5546249fb10 100644
>> --- a/net/netlink/diag.c
>> +++ b/net/netlink/diag.c
>> @@ -178,11 +178,8 @@ static int netlink_diag_dump(struct sk_buff *skb, struct netlink_callback *cb)
>> }
>> cb->args[1] = i;
>> } else {
>> - if (req->sdiag_protocol >= MAX_LINKS) {
>> - read_unlock(&nl_table_lock);
>> - rcu_read_unlock();
>> + if (req->sdiag_protocol >= MAX_LINKS)
>> return -ENOENT;
>> - }
>>
>> err = __netlink_diag_dump(skb, cb, req->sdiag_protocol, s_num);
>> }
>>
>>
^ permalink raw reply
* [PATCH net-next resend 04/13] nfp: reuse ring helpers on .ndo_open() path
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
Ring allocation helpers encapsulate all ring allocation and
initialization steps nicely. Reuse them on .ndo_open() path.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
.../net/ethernet/netronome/nfp/nfp_net_common.c | 60 ++++++++--------------
1 file changed, 20 insertions(+), 40 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index b7b2851ebb6b..50aeaea9e318 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -2051,6 +2051,13 @@ static void nfp_net_open_stack(struct nfp_net *nn)
static int nfp_net_netdev_open(struct net_device *netdev)
{
struct nfp_net *nn = netdev_priv(netdev);
+ struct nfp_net_ring_set rx = {
+ .mtu = nn->netdev->mtu,
+ .dcnt = nn->rxd_cnt,
+ };
+ struct nfp_net_ring_set tx = {
+ .dcnt = nn->txd_cnt,
+ };
int err, r;
if (nn->ctrl & NFP_NET_CFG_CTRL_ENABLE) {
@@ -2075,38 +2082,22 @@ static int nfp_net_netdev_open(struct net_device *netdev)
goto err_free_exn;
disable_irq(nn->irq_entries[NFP_NET_IRQ_LSC_IDX].vector);
- nn->rx_rings = kcalloc(nn->num_rx_rings, sizeof(*nn->rx_rings),
- GFP_KERNEL);
- if (!nn->rx_rings) {
- err = -ENOMEM;
- goto err_free_lsc;
- }
- nn->tx_rings = kcalloc(nn->num_tx_rings, sizeof(*nn->tx_rings),
- GFP_KERNEL);
- if (!nn->tx_rings) {
- err = -ENOMEM;
- goto err_free_rx_rings;
- }
-
for (r = 0; r < nn->num_r_vecs; r++) {
err = nfp_net_prepare_vector(nn, &nn->r_vecs[r], r);
if (err)
goto err_cleanup_vec_p;
}
- for (r = 0; r < nn->num_tx_rings; r++) {
- err = nfp_net_tx_ring_alloc(nn->r_vecs[r].tx_ring, nn->txd_cnt);
- if (err)
- goto err_free_tx_ring_p;
+
+ nn->rx_rings = nfp_net_rx_ring_set_prepare(nn, &rx);
+ if (!nn->rx_rings) {
+ err = -ENOMEM;
+ goto err_cleanup_vec;
}
- for (r = 0; r < nn->num_rx_rings; r++) {
- err = nfp_net_rx_ring_alloc(nn->r_vecs[r].rx_ring,
- nn->fl_bufsz, nn->rxd_cnt);
- if (err)
- goto err_flush_free_rx_ring_p;
- err = nfp_net_rx_ring_bufs_alloc(nn, nn->r_vecs[r].rx_ring);
- if (err)
- goto err_free_rx_ring_p;
+ nn->tx_rings = nfp_net_tx_ring_set_prepare(nn, &tx);
+ if (!nn->tx_rings) {
+ err = -ENOMEM;
+ goto err_free_rx_rings;
}
err = netif_set_real_num_tx_queues(netdev, nn->num_tx_rings);
@@ -2139,25 +2130,14 @@ static int nfp_net_netdev_open(struct net_device *netdev)
return 0;
err_free_rings:
- r = nn->num_rx_rings;
-err_flush_free_rx_ring_p:
- while (r--) {
- nfp_net_rx_ring_bufs_free(nn, nn->r_vecs[r].rx_ring);
-err_free_rx_ring_p:
- nfp_net_rx_ring_free(nn->r_vecs[r].rx_ring);
- }
- r = nn->num_tx_rings;
-err_free_tx_ring_p:
- while (r--)
- nfp_net_tx_ring_free(nn->r_vecs[r].tx_ring);
+ nfp_net_tx_ring_set_free(nn, &tx);
+err_free_rx_rings:
+ nfp_net_rx_ring_set_free(nn, &rx);
+err_cleanup_vec:
r = nn->num_r_vecs;
err_cleanup_vec_p:
while (r--)
nfp_net_cleanup_vector(nn, &nn->r_vecs[r]);
- kfree(nn->tx_rings);
-err_free_rx_rings:
- kfree(nn->rx_rings);
-err_free_lsc:
nfp_net_aux_irq_free(nn, NFP_NET_CFG_LSC, NFP_NET_IRQ_LSC_IDX);
err_free_exn:
nfp_net_aux_irq_free(nn, NFP_NET_CFG_EXN, NFP_NET_IRQ_EXN_IDX);
--
1.9.1
^ permalink raw reply related
* Re: [PATCH net-next v2 1/1] driver: veth: Refine the statistics codes of veth driver
From: David Miller @ 2016-11-03 17:18 UTC (permalink / raw)
To: fgao; +Cc: eric.dumazet, cwang, vijayp, ej, pabeni, netdev
In-Reply-To: <CA+6hz4owi2WgeGWd98=8d=Q5DBPY4_ZtYAOgcM1V_0+OU8UHrw@mail.gmail.com>
From: Gao Feng <fgao@ikuai8.com>
Date: Thu, 3 Nov 2016 22:38:28 +0800
> On Thu, Nov 3, 2016 at 10:31 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
>> On Thu, 2016-11-03 at 21:39 +0800, Gao Feng wrote:
>>> Hi Eric,
>>>
>>> On Thu, Nov 3, 2016 at 9:30 PM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
>>> > On Thu, 2016-11-03 at 21:03 +0800, fgao@ikuai8.com wrote:
>>> >> From: Gao Feng <fgao@ikuai8.com>
>>> >>
>>> >> The dropped count of veth is located in struct veth_priv, but other
>>> >> statistics like packets and bytes are in another struct pcpu_vstats.
>>> >> Now keep these three counters in the same struct.
>>> >>
>>> >> Signed-off-by: Gao Feng <fgao@ikuai8.com>
>>> >> ---
>>> >> v2: Use right "peer" instead of "dev";
>>> >> v1: Initial version
>>> >
>>> > May I ask : Why ?
>>>
>>> Just because I think statistics should be in the same struct.
>>
>> That is not a good reason then.
>
> Because other net devices put the statistics together.
Organizational "prettyness" is not argument for this change, when the
downsides are fundamentally clear:
1) It is not a fast-path accessed statistic, so the per-cpu'ness is
not important.
2) We aim to minimize the amount of per-cpu data in the kernel because
it is expensive. So when not necessary, as is the case here, we
do not user per-cpu data.
There are no good reasons to make this change, so I am dropping your
patch.
^ permalink raw reply
* [PATCH net-next resend 03/13] nfp: rename ring allocation helpers
From: Jakub Kicinski @ 2016-11-03 17:11 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
"Shadow" in ring helpers used to mean that the helper will allocate
rings without touching existing configuration, this was used for
reconfiguration while the device was running. We will soon use
the same helpers for .ndo_open() path, so replace "shadow" with
"ring_set".
No functional changes.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
.../net/ethernet/netronome/nfp/nfp_net_common.c | 26 +++++++++++-----------
1 file changed, 13 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index e58532d27c5b..b7b2851ebb6b 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -1573,7 +1573,7 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
}
static struct nfp_net_tx_ring *
-nfp_net_shadow_tx_rings_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_tx_ring_set_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
struct nfp_net_tx_ring *rings;
unsigned int r;
@@ -1599,7 +1599,7 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
}
static void
-nfp_net_shadow_tx_rings_swap(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_tx_ring_set_swap(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
struct nfp_net_tx_ring *rings = s->rings;
struct nfp_net_ring_set new = *s;
@@ -1616,7 +1616,7 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
}
static void
-nfp_net_shadow_tx_rings_free(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_tx_ring_set_free(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
struct nfp_net_tx_ring *rings = s->rings;
unsigned int r;
@@ -1693,7 +1693,7 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
}
static struct nfp_net_rx_ring *
-nfp_net_shadow_rx_rings_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_rx_ring_set_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
unsigned int fl_bufsz = nfp_net_calc_fl_bufsz(nn, s->mtu);
struct nfp_net_rx_ring *rings;
@@ -1726,7 +1726,7 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
}
static void
-nfp_net_shadow_rx_rings_swap(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_rx_ring_set_swap(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
struct nfp_net_rx_ring *rings = s->rings;
struct nfp_net_ring_set new = *s;
@@ -1746,7 +1746,7 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
}
static void
-nfp_net_shadow_rx_rings_free(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_rx_ring_set_free(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
struct nfp_net_rx_ring *rings = s->rings;
unsigned int r;
@@ -2268,9 +2268,9 @@ static void nfp_net_set_rx_mode(struct net_device *netdev)
struct nfp_net_ring_set *tx)
{
if (rx)
- nfp_net_shadow_rx_rings_swap(nn, rx);
+ nfp_net_rx_ring_set_swap(nn, rx);
if (tx)
- nfp_net_shadow_tx_rings_swap(nn, tx);
+ nfp_net_tx_ring_set_swap(nn, tx);
return __nfp_net_set_config_and_enable(nn);
}
@@ -2299,11 +2299,11 @@ static void nfp_net_set_rx_mode(struct net_device *netdev)
/* Prepare new rings */
if (rx) {
- if (!nfp_net_shadow_rx_rings_prepare(nn, rx))
+ if (!nfp_net_rx_ring_set_prepare(nn, rx))
return -ENOMEM;
}
if (tx) {
- if (!nfp_net_shadow_tx_rings_prepare(nn, tx)) {
+ if (!nfp_net_tx_ring_set_prepare(nn, tx)) {
err = -ENOMEM;
goto err_free_rx;
}
@@ -2327,9 +2327,9 @@ static void nfp_net_set_rx_mode(struct net_device *netdev)
}
if (rx)
- nfp_net_shadow_rx_rings_free(nn, rx);
+ nfp_net_rx_ring_set_free(nn, rx);
if (tx)
- nfp_net_shadow_tx_rings_free(nn, tx);
+ nfp_net_tx_ring_set_free(nn, tx);
nfp_net_open_stack(nn);
@@ -2337,7 +2337,7 @@ static void nfp_net_set_rx_mode(struct net_device *netdev)
err_free_rx:
if (rx)
- nfp_net_shadow_rx_rings_free(nn, rx);
+ nfp_net_rx_ring_set_free(nn, rx);
return err;
}
--
1.9.1
^ permalink raw reply related
* Re: [PATCH net-next v1 00/21] amd-xgbe: AMD XGBE driver updates 2016-11-01
From: David Miller @ 2016-11-03 17:14 UTC (permalink / raw)
To: thomas.lendacky; +Cc: netdev, f.fainelli
In-Reply-To: <20161103132744.3437.27332.stgit@tlendack-t1.amdoffice.net>
20+ patches is too many to submit at one time.
Making huge patch series submissions puts an unreasonable burdon on
those developers who might decide to review your work, including me.
Please keep your series down to a small, reasonable, size. Perhaps
10-15 patches maximum.
I will not be looking at nor applying this series.
^ permalink raw reply
* [PATCH] net: icmp_route_lookup should use rt dev to determine L3 domain
From: David Ahern @ 2016-11-03 17:13 UTC (permalink / raw)
To: netdev; +Cc: David Ahern
icmp_send is called in response to some event. The skb may not have
the device set (skb->dev is NULL), but it is expected to have an rt.
Update icmp_route_lookup to use the rt on the skb to determine L3
domain.
Fixes: 613d09b30f8b ("net: Use VRF device index for lookups on TX")
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
---
net/ipv4/icmp.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 38abe70e595f..774a15e9f041 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -477,7 +477,7 @@ static struct rtable *icmp_route_lookup(struct net *net,
fl4->flowi4_proto = IPPROTO_ICMP;
fl4->fl4_icmp_type = type;
fl4->fl4_icmp_code = code;
- fl4->flowi4_oif = l3mdev_master_ifindex(skb_in->dev);
+ fl4->flowi4_oif = l3mdev_master_ifindex(skb_rtable(skb_in)->dst.dev);
security_skb_classify_flow(skb_in, flowi4_to_flowi(fl4));
rt = __ip_route_output_key_hash(net, fl4,
@@ -502,7 +502,7 @@ static struct rtable *icmp_route_lookup(struct net *net,
if (err)
goto relookup_failed;
- if (inet_addr_type_dev_table(net, skb_in->dev,
+ if (inet_addr_type_dev_table(net, skb_rtable(skb_in)->dst.dev,
fl4_dec.saddr) == RTN_LOCAL) {
rt2 = __ip_route_output_key(net, &fl4_dec);
if (IS_ERR(rt2))
--
2.1.4
^ permalink raw reply related
* [PATCH net-next resend 13/13] nfp: add support for offload of XDP programs
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
Most infrastructure can be reused, provide separate handling
of context offsets and exit codes.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/nfp_bpf.h | 1 +
drivers/net/ethernet/netronome/nfp/nfp_bpf_jit.c | 92 +++++++++++++++++++++-
.../net/ethernet/netronome/nfp/nfp_bpf_verifier.c | 3 +
drivers/net/ethernet/netronome/nfp/nfp_net.h | 2 +
.../net/ethernet/netronome/nfp/nfp_net_common.c | 44 ++++++++++-
.../net/ethernet/netronome/nfp/nfp_net_offload.c | 3 +
6 files changed, 139 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_bpf.h b/drivers/net/ethernet/netronome/nfp/nfp_bpf.h
index 87aa8a3e9112..76a19f1796af 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_bpf.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_bpf.h
@@ -62,6 +62,7 @@ enum nfp_bpf_action_type {
NN_ACT_TC_DROP,
NN_ACT_TC_REDIR,
NN_ACT_DIRECT,
+ NN_ACT_XDP,
};
/* Software register representation, hardware encoding in asm.h */
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_bpf_jit.c b/drivers/net/ethernet/netronome/nfp/nfp_bpf_jit.c
index f8df5300f49c..335beb8b8b45 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_bpf_jit.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_bpf_jit.c
@@ -1126,7 +1126,7 @@ static int data_ind_ld4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
meta->insn.src_reg * 2, true, 4);
}
-static int mem_ldx4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+static int mem_ldx4_skb(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
{
if (meta->insn.off == offsetof(struct sk_buff, len))
emit_alu(nfp_prog, reg_both(meta->insn.dst_reg * 2),
@@ -1134,12 +1134,42 @@ static int mem_ldx4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
else
return -ENOTSUPP;
- wrp_immed(nfp_prog, reg_both(meta->insn.dst_reg * 2 + 1), 0);
+ return 0;
+}
+
+static int mem_ldx4_xdp(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+{
+ u32 dst = reg_both(meta->insn.dst_reg * 2);
+
+ if (meta->insn.off != offsetof(struct xdp_md, data) &&
+ meta->insn.off != offsetof(struct xdp_md, data_end))
+ return -ENOTSUPP;
+
+ emit_alu(nfp_prog, dst, reg_none(), ALU_OP_NONE, NFP_BPF_ABI_PKT);
+
+ if (meta->insn.off == offsetof(struct xdp_md, data))
+ return 0;
+
+ emit_alu(nfp_prog, dst, dst, ALU_OP_ADD, NFP_BPF_ABI_LEN);
return 0;
}
-static int mem_stx4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+static int mem_ldx4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+{
+ int ret;
+
+ if (nfp_prog->act == NN_ACT_XDP)
+ ret = mem_ldx4_xdp(nfp_prog, meta);
+ else
+ ret = mem_ldx4_skb(nfp_prog, meta);
+
+ wrp_immed(nfp_prog, reg_both(meta->insn.dst_reg * 2 + 1), 0);
+
+ return ret;
+}
+
+static int mem_stx4_skb(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
{
if (meta->insn.off == offsetof(struct sk_buff, mark))
return wrp_set_mark(nfp_prog, meta->insn.src_reg * 2);
@@ -1147,6 +1177,18 @@ static int mem_stx4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
return -ENOTSUPP;
}
+static int mem_stx4_xdp(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+{
+ return -ENOTSUPP;
+}
+
+static int mem_stx4(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
+{
+ if (nfp_prog->act == NN_ACT_XDP)
+ return mem_stx4_xdp(nfp_prog, meta);
+ return mem_stx4_skb(nfp_prog, meta);
+}
+
static int jump(struct nfp_prog *nfp_prog, struct nfp_insn_meta *meta)
{
if (meta->insn.off < 0) /* TODO */
@@ -1530,6 +1572,47 @@ static void nfp_outro_tc_da(struct nfp_prog *nfp_prog)
emit_ld_field(nfp_prog, reg_a(0), 0xc, reg_b(2), SHF_SC_L_SHF, 16);
}
+static void nfp_outro_xdp(struct nfp_prog *nfp_prog)
+{
+ /* XDP return codes:
+ * 0 aborted 0x82 -> drop, count as stat3
+ * 1 drop 0x22 -> drop, count as stat1
+ * 2 pass 0x11 -> pass, count as stat0
+ * 3 tx 0x44 -> redir, count as stat2
+ * * unknown 0x82 -> drop, count as stat3
+ */
+ /* Target for aborts */
+ nfp_prog->tgt_abort = nfp_prog_current_offset(nfp_prog);
+
+ emit_br_def(nfp_prog, nfp_prog->tgt_done, 2);
+
+ emit_alu(nfp_prog, reg_a(0),
+ reg_none(), ALU_OP_NONE, NFP_BPF_ABI_FLAGS);
+ emit_ld_field(nfp_prog, reg_a(0), 0xc, reg_imm(0x82), SHF_SC_L_SHF, 16);
+
+ /* Target for normal exits */
+ nfp_prog->tgt_out = nfp_prog_current_offset(nfp_prog);
+
+ /* if R0 > 3 jump to abort */
+ emit_alu(nfp_prog, reg_none(), reg_imm(3), ALU_OP_SUB, reg_b(0));
+ emit_br(nfp_prog, BR_BLO, nfp_prog->tgt_abort, 0);
+
+ wrp_immed(nfp_prog, reg_b(2), 0x44112282);
+
+ emit_shf(nfp_prog, reg_a(1),
+ reg_none(), SHF_OP_NONE, reg_b(0), SHF_SC_L_SHF, 3);
+
+ emit_alu(nfp_prog, reg_none(), reg_a(1), ALU_OP_OR, reg_imm(0));
+ emit_shf(nfp_prog, reg_b(2),
+ reg_imm(0xff), SHF_OP_AND, reg_b(2), SHF_SC_R_SHF, 0);
+
+ emit_br_def(nfp_prog, nfp_prog->tgt_done, 2);
+
+ emit_alu(nfp_prog, reg_a(0),
+ reg_none(), ALU_OP_NONE, NFP_BPF_ABI_FLAGS);
+ emit_ld_field(nfp_prog, reg_a(0), 0xc, reg_b(2), SHF_SC_L_SHF, 16);
+}
+
static void nfp_outro(struct nfp_prog *nfp_prog)
{
switch (nfp_prog->act) {
@@ -1540,6 +1623,9 @@ static void nfp_outro(struct nfp_prog *nfp_prog)
case NN_ACT_TC_REDIR:
nfp_outro_tc_legacy(nfp_prog);
break;
+ case NN_ACT_XDP:
+ nfp_outro_xdp(nfp_prog);
+ break;
}
}
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_bpf_verifier.c b/drivers/net/ethernet/netronome/nfp/nfp_bpf_verifier.c
index 144cae87f63a..b3361f9b8e5c 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_bpf_verifier.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_bpf_verifier.c
@@ -80,6 +80,9 @@ struct nfp_bpf_analyzer_priv {
{
const struct bpf_reg_state *reg0 = &env->cur_state.regs[0];
+ if (nfp_prog->act == NN_ACT_XDP)
+ return 0;
+
if (reg0->type != CONST_IMM) {
pr_info("unsupported exit state: %d, imm: %llx\n",
reg0->type, reg0->imm);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
index fd29a6306991..2115f446031e 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
@@ -435,6 +435,7 @@ struct nfp_stat_pair {
* @is_vf: Is the driver attached to a VF?
* @fw_loaded: Is the firmware loaded?
* @bpf_offload_skip_sw: Offloaded BPF program will not be rerun by cls_bpf
+ * @bpf_offload_xdp: Offloaded BPF program is XDP
* @ctrl: Local copy of the control register/word.
* @fl_bufsz: Currently configured size of the freelist buffers
* @rx_offset: Offset in the RX buffers where packet data starts
@@ -502,6 +503,7 @@ struct nfp_net {
unsigned is_vf:1;
unsigned fw_loaded:1;
unsigned bpf_offload_skip_sw:1;
+ unsigned bpf_offload_xdp:1;
u32 ctrl;
u32 fl_bufsz;
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 1e8e00d25c51..99edb9fd84bf 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -1600,7 +1600,8 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
r_vec->rx_bytes += pkt_len;
u64_stats_update_end(&r_vec->rx_sync);
- if (xdp_prog) {
+ if (xdp_prog && !(rxd->rxd.flags & PCIE_DESC_RX_BPF &&
+ nn->bpf_offload_xdp)) {
int act;
dma_sync_single_for_cpu(&nn->pdev->dev,
@@ -2693,8 +2694,12 @@ static bool nfp_net_ebpf_capable(struct nfp_net *nn)
if (proto != htons(ETH_P_ALL))
return -ENOTSUPP;
- if (tc->type == TC_SETUP_CLSBPF && nfp_net_ebpf_capable(nn))
- return nfp_net_bpf_offload(nn, tc->cls_bpf);
+ if (tc->type == TC_SETUP_CLSBPF && nfp_net_ebpf_capable(nn)) {
+ if (!nn->bpf_offload_xdp)
+ return nfp_net_bpf_offload(nn, tc->cls_bpf);
+ else
+ return -EBUSY;
+ }
return -EINVAL;
}
@@ -2902,6 +2907,34 @@ static void nfp_net_del_vxlan_port(struct net_device *netdev,
nfp_net_set_vxlan_port(nn, idx, 0);
}
+static int nfp_net_xdp_offload(struct nfp_net *nn, struct bpf_prog *prog)
+{
+ struct tc_cls_bpf_offload cmd = {
+ .prog = prog,
+ };
+ int ret;
+
+ if (!nfp_net_ebpf_capable(nn))
+ return -EINVAL;
+
+ if (nn->ctrl & NFP_NET_CFG_CTRL_BPF) {
+ if (!nn->bpf_offload_xdp)
+ return prog ? -EBUSY : 0;
+ cmd.command = prog ? TC_CLSBPF_REPLACE : TC_CLSBPF_DESTROY;
+ } else {
+ if (!prog)
+ return 0;
+ cmd.command = TC_CLSBPF_ADD;
+ }
+
+ ret = nfp_net_bpf_offload(nn, &cmd);
+ /* Stop offload if replace not possible */
+ if (ret && cmd.command == TC_CLSBPF_REPLACE)
+ nfp_net_xdp_offload(nn, NULL);
+ nn->bpf_offload_xdp = prog && !ret;
+ return ret;
+}
+
static int nfp_net_xdp_setup(struct nfp_net *nn, struct bpf_prog *prog)
{
struct nfp_net_ring_set rx = {
@@ -2920,6 +2953,7 @@ static int nfp_net_xdp_setup(struct nfp_net *nn, struct bpf_prog *prog)
if (prog && nn->xdp_prog) {
prog = xchg(&nn->xdp_prog, prog);
bpf_prog_put(prog);
+ nfp_net_xdp_offload(nn, nn->xdp_prog);
return 0;
}
@@ -2934,6 +2968,8 @@ static int nfp_net_xdp_setup(struct nfp_net *nn, struct bpf_prog *prog)
if (prog)
bpf_prog_put(prog);
+ nfp_net_xdp_offload(nn, nn->xdp_prog);
+
return 0;
}
@@ -3233,5 +3269,7 @@ void nfp_net_netdev_clean(struct net_device *netdev)
if (nn->xdp_prog)
bpf_prog_put(nn->xdp_prog);
+ if (nn->bpf_offload_xdp)
+ nfp_net_xdp_offload(nn, NULL);
unregister_netdev(nn->netdev);
}
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c b/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c
index 4bb6f16e2a7a..18a851eb3508 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c
@@ -111,6 +111,9 @@ static void nfp_net_bpf_stats_reset(struct nfp_net *nn)
const struct tc_action *a;
LIST_HEAD(actions);
+ if (!cls_bpf->exts)
+ return NN_ACT_XDP;
+
/* TC direct action */
if (cls_bpf->exts_integrated) {
if (tc_no_actions(cls_bpf->exts))
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 11/13] nfp: add XDP support in the driver
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
Add XDP support. Separate stack's and XDP's TX rings logically.
Add functions for handling XDP_TX and cleanup of XDP's TX rings.
For XDP allocate all RX buffers as separate pages and map them
with DMA_BIDIRECTIONAL.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/nfp_net.h | 17 +-
.../net/ethernet/netronome/nfp/nfp_net_common.c | 474 +++++++++++++++++----
.../net/ethernet/netronome/nfp/nfp_net_debugfs.c | 37 +-
.../net/ethernet/netronome/nfp/nfp_net_ethtool.c | 22 +-
4 files changed, 449 insertions(+), 101 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
index 486e7c6453bc..abc9e56e93b8 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
@@ -171,7 +171,10 @@ struct nfp_net_tx_desc {
* on the head's buffer). Equal to skb->len for non-TSO packets.
*/
struct nfp_net_tx_buf {
- struct sk_buff *skb;
+ union {
+ struct sk_buff *skb;
+ void *frag;
+ };
dma_addr_t dma_addr;
short int fidx;
u16 pkt_cnt;
@@ -341,6 +344,7 @@ struct nfp_net_rx_ring {
* @napi: NAPI structure for this ring vec
* @tx_ring: Pointer to TX ring
* @rx_ring: Pointer to RX ring
+ * @xdp_ring: Pointer to an extra TX ring for XDP
* @irq_idx: Index into MSI-X table
* @rx_sync: Seqlock for atomic updates of RX stats
* @rx_pkts: Number of received packets
@@ -384,6 +388,8 @@ struct nfp_net_r_vector {
u64 hw_csum_rx_inner_ok;
u64 hw_csum_rx_error;
+ struct nfp_net_tx_ring *xdp_ring;
+
struct u64_stats_sync tx_sync;
u64 tx_pkts;
u64 tx_bytes;
@@ -432,6 +438,7 @@ struct nfp_stat_pair {
* @ctrl: Local copy of the control register/word.
* @fl_bufsz: Currently configured size of the freelist buffers
* @rx_offset: Offset in the RX buffers where packet data starts
+ * @xdp_prog: Installed XDP program
* @cpp: Pointer to the CPP handle
* @nfp_dev_cpp: Pointer to the NFP Device handle
* @ctrl_area: Pointer to the CPP area for the control BAR
@@ -451,6 +458,7 @@ struct nfp_stat_pair {
* @max_tx_rings: Maximum number of TX rings supported by the Firmware
* @max_rx_rings: Maximum number of RX rings supported by the Firmware
* @num_tx_rings: Currently configured number of TX rings
+ * @num_stack_tx_rings: Number of TX rings used by the stack (not XDP)
* @num_rx_rings: Currently configured number of RX rings
* @txd_cnt: Size of the TX ring in number of descriptors
* @rxd_cnt: Size of the RX ring in number of descriptors
@@ -500,6 +508,8 @@ struct nfp_net {
u32 rx_offset;
+ struct bpf_prog *xdp_prog;
+
struct nfp_net_tx_ring *tx_rings;
struct nfp_net_rx_ring *rx_rings;
@@ -532,6 +542,7 @@ struct nfp_net {
unsigned int max_rx_rings;
unsigned int num_tx_rings;
+ unsigned int num_stack_tx_rings;
unsigned int num_rx_rings;
int stride_tx;
@@ -779,8 +790,8 @@ struct nfp_net *
int nfp_net_irqs_alloc(struct nfp_net *nn);
void nfp_net_irqs_disable(struct nfp_net *nn);
int
-nfp_net_ring_reconfig(struct nfp_net *nn, struct nfp_net_ring_set *rx,
- struct nfp_net_ring_set *tx);
+nfp_net_ring_reconfig(struct nfp_net *nn, struct bpf_prog **xdp_prog,
+ struct nfp_net_ring_set *rx, struct nfp_net_ring_set *tx);
#ifdef CONFIG_NFP_NET_DEBUG
void nfp_net_debugfs_create(void);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 2ab63661a6fd..fa43dbcecc4f 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -41,6 +41,7 @@
* Chris Telfer <chris.telfer@netronome.com>
*/
+#include <linux/bpf.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/init.h>
@@ -490,6 +491,7 @@ static void nfp_net_irqs_assign(struct net_device *netdev)
nn->num_rx_rings = min(nn->num_r_vecs, nn->num_rx_rings);
nn->num_tx_rings = min(nn->num_r_vecs, nn->num_tx_rings);
+ nn->num_stack_tx_rings = nn->num_tx_rings;
nn->lsc_handler = nfp_net_irq_lsc;
nn->exn_handler = nfp_net_irq_exn;
@@ -713,6 +715,13 @@ static void nfp_net_tx_csum(struct nfp_net *nn, struct nfp_net_r_vector *r_vec,
u64_stats_update_end(&r_vec->tx_sync);
}
+static void nfp_net_tx_xmit_more_flush(struct nfp_net_tx_ring *tx_ring)
+{
+ wmb();
+ nfp_qcp_wr_ptr_add(tx_ring->qcp_q, tx_ring->wr_ptr_add);
+ tx_ring->wr_ptr_add = 0;
+}
+
/**
* nfp_net_tx() - Main transmit entry point
* @skb: SKB to transmit
@@ -827,12 +836,8 @@ static int nfp_net_tx(struct sk_buff *skb, struct net_device *netdev)
nfp_net_tx_ring_stop(nd_q, tx_ring);
tx_ring->wr_ptr_add += nr_frags + 1;
- if (!skb->xmit_more || netif_xmit_stopped(nd_q)) {
- /* force memory write before we let HW know */
- wmb();
- nfp_qcp_wr_ptr_add(tx_ring->qcp_q, tx_ring->wr_ptr_add);
- tx_ring->wr_ptr_add = 0;
- }
+ if (!skb->xmit_more || netif_xmit_stopped(nd_q))
+ nfp_net_tx_xmit_more_flush(tx_ring);
skb_tx_timestamp(skb);
@@ -954,6 +959,56 @@ static void nfp_net_tx_complete(struct nfp_net_tx_ring *tx_ring)
tx_ring->rd_p, tx_ring->wr_p, tx_ring->cnt);
}
+static void nfp_net_xdp_complete(struct nfp_net_tx_ring *tx_ring)
+{
+ struct nfp_net_r_vector *r_vec = tx_ring->r_vec;
+ struct nfp_net *nn = r_vec->nfp_net;
+ u32 done_pkts = 0, done_bytes = 0;
+ int idx, todo;
+ u32 qcp_rd_p;
+
+ /* Work out how many descriptors have been transmitted */
+ qcp_rd_p = nfp_qcp_rd_ptr_read(tx_ring->qcp_q);
+
+ if (qcp_rd_p == tx_ring->qcp_rd_p)
+ return;
+
+ if (qcp_rd_p > tx_ring->qcp_rd_p)
+ todo = qcp_rd_p - tx_ring->qcp_rd_p;
+ else
+ todo = qcp_rd_p + tx_ring->cnt - tx_ring->qcp_rd_p;
+
+ while (todo--) {
+ idx = tx_ring->rd_p & (tx_ring->cnt - 1);
+ tx_ring->rd_p++;
+
+ if (!tx_ring->txbufs[idx].frag)
+ continue;
+
+ nfp_net_dma_unmap_rx(nn, tx_ring->txbufs[idx].dma_addr,
+ nn->fl_bufsz, DMA_BIDIRECTIONAL);
+ __free_page(virt_to_page(tx_ring->txbufs[idx].frag));
+
+ done_pkts++;
+ done_bytes += tx_ring->txbufs[idx].real_len;
+
+ tx_ring->txbufs[idx].dma_addr = 0;
+ tx_ring->txbufs[idx].frag = NULL;
+ tx_ring->txbufs[idx].fidx = -2;
+ }
+
+ tx_ring->qcp_rd_p = qcp_rd_p;
+
+ u64_stats_update_begin(&r_vec->tx_sync);
+ r_vec->tx_bytes += done_bytes;
+ r_vec->tx_pkts += done_pkts;
+ u64_stats_update_end(&r_vec->tx_sync);
+
+ WARN_ONCE(tx_ring->wr_p - tx_ring->rd_p > tx_ring->cnt,
+ "TX ring corruption rd_p=%u wr_p=%u cnt=%u\n",
+ tx_ring->rd_p, tx_ring->wr_p, tx_ring->cnt);
+}
+
/**
* nfp_net_tx_ring_reset() - Free any untransmitted buffers and reset pointers
* @nn: NFP Net device
@@ -964,39 +1019,47 @@ static void nfp_net_tx_complete(struct nfp_net_tx_ring *tx_ring)
static void
nfp_net_tx_ring_reset(struct nfp_net *nn, struct nfp_net_tx_ring *tx_ring)
{
+ struct nfp_net_r_vector *r_vec = tx_ring->r_vec;
const struct skb_frag_struct *frag;
- struct netdev_queue *nd_q;
struct pci_dev *pdev = nn->pdev;
+ struct netdev_queue *nd_q;
while (tx_ring->rd_p != tx_ring->wr_p) {
- int nr_frags, fidx, idx;
- struct sk_buff *skb;
+ struct nfp_net_tx_buf *tx_buf;
+ int idx;
idx = tx_ring->rd_p & (tx_ring->cnt - 1);
- skb = tx_ring->txbufs[idx].skb;
- nr_frags = skb_shinfo(skb)->nr_frags;
- fidx = tx_ring->txbufs[idx].fidx;
+ tx_buf = &tx_ring->txbufs[idx];
- if (fidx == -1) {
- /* unmap head */
- dma_unmap_single(&pdev->dev,
- tx_ring->txbufs[idx].dma_addr,
- skb_headlen(skb), DMA_TO_DEVICE);
+ if (tx_ring == r_vec->xdp_ring) {
+ nfp_net_dma_unmap_rx(nn, tx_buf->dma_addr,
+ nn->fl_bufsz, DMA_BIDIRECTIONAL);
+ __free_page(virt_to_page(tx_ring->txbufs[idx].frag));
} else {
- /* unmap fragment */
- frag = &skb_shinfo(skb)->frags[fidx];
- dma_unmap_page(&pdev->dev,
- tx_ring->txbufs[idx].dma_addr,
- skb_frag_size(frag), DMA_TO_DEVICE);
- }
+ struct sk_buff *skb = tx_ring->txbufs[idx].skb;
+ int nr_frags = skb_shinfo(skb)->nr_frags;
+
+ if (tx_buf->fidx == -1) {
+ /* unmap head */
+ dma_unmap_single(&pdev->dev, tx_buf->dma_addr,
+ skb_headlen(skb),
+ DMA_TO_DEVICE);
+ } else {
+ /* unmap fragment */
+ frag = &skb_shinfo(skb)->frags[tx_buf->fidx];
+ dma_unmap_page(&pdev->dev, tx_buf->dma_addr,
+ skb_frag_size(frag),
+ DMA_TO_DEVICE);
+ }
- /* check for last gather fragment */
- if (fidx == nr_frags - 1)
- dev_kfree_skb_any(skb);
+ /* check for last gather fragment */
+ if (tx_buf->fidx == nr_frags - 1)
+ dev_kfree_skb_any(skb);
+ }
- tx_ring->txbufs[idx].dma_addr = 0;
- tx_ring->txbufs[idx].skb = NULL;
- tx_ring->txbufs[idx].fidx = -2;
+ tx_buf->dma_addr = 0;
+ tx_buf->skb = NULL;
+ tx_buf->fidx = -2;
tx_ring->qcp_rd_p++;
tx_ring->rd_p++;
@@ -1008,6 +1071,9 @@ static void nfp_net_tx_complete(struct nfp_net_tx_ring *tx_ring)
tx_ring->qcp_rd_p = 0;
tx_ring->wr_ptr_add = 0;
+ if (tx_ring == r_vec->xdp_ring)
+ return;
+
nd_q = netdev_get_tx_queue(nn->netdev, tx_ring->idx);
netdev_tx_reset_queue(nd_q);
}
@@ -1017,7 +1083,7 @@ static void nfp_net_tx_timeout(struct net_device *netdev)
struct nfp_net *nn = netdev_priv(netdev);
int i;
- for (i = 0; i < nn->num_tx_rings; i++) {
+ for (i = 0; i < nn->netdev->real_num_tx_queues; i++) {
if (!netif_tx_queue_stopped(netdev_get_tx_queue(netdev, i)))
continue;
nn_warn(nn, "TX timeout on ring: %d\n", i);
@@ -1045,11 +1111,21 @@ static void nfp_net_tx_timeout(struct net_device *netdev)
return fl_bufsz;
}
+static void
+nfp_net_free_frag(void *frag, bool xdp)
+{
+ if (!xdp)
+ skb_free_frag(frag);
+ else
+ __free_page(virt_to_page(frag));
+}
+
/**
* nfp_net_rx_alloc_one() - Allocate and map page frag for RX
* @rx_ring: RX ring structure of the skb
* @dma_addr: Pointer to storage for DMA address (output param)
* @fl_bufsz: size of freelist buffers
+ * @xdp: Whether XDP is enabled
*
* This function will allcate a new page frag, map it for DMA.
*
@@ -1057,20 +1133,26 @@ static void nfp_net_tx_timeout(struct net_device *netdev)
*/
static void *
nfp_net_rx_alloc_one(struct nfp_net_rx_ring *rx_ring, dma_addr_t *dma_addr,
- unsigned int fl_bufsz)
+ unsigned int fl_bufsz, bool xdp)
{
struct nfp_net *nn = rx_ring->r_vec->nfp_net;
+ int direction;
void *frag;
- frag = netdev_alloc_frag(fl_bufsz);
+ if (!xdp)
+ frag = netdev_alloc_frag(fl_bufsz);
+ else
+ frag = page_address(alloc_page(GFP_KERNEL | __GFP_COLD));
if (!frag) {
nn_warn_ratelimit(nn, "Failed to alloc receive page frag\n");
return NULL;
}
- *dma_addr = nfp_net_dma_map_rx(nn, frag, fl_bufsz, DMA_FROM_DEVICE);
+ direction = xdp ? DMA_BIDIRECTIONAL : DMA_FROM_DEVICE;
+
+ *dma_addr = nfp_net_dma_map_rx(nn, frag, fl_bufsz, direction);
if (dma_mapping_error(&nn->pdev->dev, *dma_addr)) {
- skb_free_frag(frag);
+ nfp_net_free_frag(frag, xdp);
nn_warn_ratelimit(nn, "Failed to map DMA RX buffer\n");
return NULL;
}
@@ -1078,19 +1160,23 @@ static void nfp_net_tx_timeout(struct net_device *netdev)
return frag;
}
-static void *nfp_net_napi_alloc_one(struct nfp_net *nn, dma_addr_t *dma_addr)
+static void *
+nfp_net_napi_alloc_one(struct nfp_net *nn, int direction, dma_addr_t *dma_addr)
{
void *frag;
- frag = napi_alloc_frag(nn->fl_bufsz);
+ if (!nn->xdp_prog)
+ frag = napi_alloc_frag(nn->fl_bufsz);
+ else
+ frag = page_address(alloc_page(GFP_ATOMIC | __GFP_COLD));
if (!frag) {
nn_warn_ratelimit(nn, "Failed to alloc receive page frag\n");
return NULL;
}
- *dma_addr = nfp_net_dma_map_rx(nn, frag, nn->fl_bufsz, DMA_FROM_DEVICE);
+ *dma_addr = nfp_net_dma_map_rx(nn, frag, nn->fl_bufsz, direction);
if (dma_mapping_error(&nn->pdev->dev, *dma_addr)) {
- skb_free_frag(frag);
+ nfp_net_free_frag(frag, nn->xdp_prog);
nn_warn_ratelimit(nn, "Failed to map DMA RX buffer\n");
return NULL;
}
@@ -1161,14 +1247,17 @@ static void nfp_net_rx_ring_reset(struct nfp_net_rx_ring *rx_ring)
* nfp_net_rx_ring_bufs_free() - Free any buffers currently on the RX ring
* @nn: NFP Net device
* @rx_ring: RX ring to remove buffers from
+ * @xdp: Whether XDP is enabled
*
* Assumes that the device is stopped and buffers are in [0, ring->cnt - 1)
* entries. After device is disabled nfp_net_rx_ring_reset() must be called
* to restore required ring geometry.
*/
static void
-nfp_net_rx_ring_bufs_free(struct nfp_net *nn, struct nfp_net_rx_ring *rx_ring)
+nfp_net_rx_ring_bufs_free(struct nfp_net *nn, struct nfp_net_rx_ring *rx_ring,
+ bool xdp)
{
+ int direction = xdp ? DMA_BIDIRECTIONAL : DMA_FROM_DEVICE;
unsigned int i;
for (i = 0; i < rx_ring->cnt - 1; i++) {
@@ -1180,8 +1269,8 @@ static void nfp_net_rx_ring_reset(struct nfp_net_rx_ring *rx_ring)
continue;
nfp_net_dma_unmap_rx(nn, rx_ring->rxbufs[i].dma_addr,
- rx_ring->bufsz, DMA_FROM_DEVICE);
- skb_free_frag(rx_ring->rxbufs[i].frag);
+ rx_ring->bufsz, direction);
+ nfp_net_free_frag(rx_ring->rxbufs[i].frag, xdp);
rx_ring->rxbufs[i].dma_addr = 0;
rx_ring->rxbufs[i].frag = NULL;
}
@@ -1191,9 +1280,11 @@ static void nfp_net_rx_ring_reset(struct nfp_net_rx_ring *rx_ring)
* nfp_net_rx_ring_bufs_alloc() - Fill RX ring with buffers (don't give to FW)
* @nn: NFP Net device
* @rx_ring: RX ring to remove buffers from
+ * @xdp: Whether XDP is enabled
*/
static int
-nfp_net_rx_ring_bufs_alloc(struct nfp_net *nn, struct nfp_net_rx_ring *rx_ring)
+nfp_net_rx_ring_bufs_alloc(struct nfp_net *nn, struct nfp_net_rx_ring *rx_ring,
+ bool xdp)
{
struct nfp_net_rx_buf *rxbufs;
unsigned int i;
@@ -1203,9 +1294,9 @@ static void nfp_net_rx_ring_reset(struct nfp_net_rx_ring *rx_ring)
for (i = 0; i < rx_ring->cnt - 1; i++) {
rxbufs[i].frag =
nfp_net_rx_alloc_one(rx_ring, &rxbufs[i].dma_addr,
- rx_ring->bufsz);
+ rx_ring->bufsz, xdp);
if (!rxbufs[i].frag) {
- nfp_net_rx_ring_bufs_free(nn, rx_ring);
+ nfp_net_rx_ring_bufs_free(nn, rx_ring, xdp);
return -ENOMEM;
}
}
@@ -1368,6 +1459,68 @@ static void nfp_net_set_hash(struct net_device *netdev, struct sk_buff *skb,
dev_kfree_skb_any(skb);
}
+static void
+nfp_net_tx_xdp_buf(struct nfp_net *nn, struct nfp_net_rx_ring *rx_ring,
+ struct nfp_net_tx_ring *tx_ring,
+ struct nfp_net_rx_buf *rxbuf, unsigned int pkt_off,
+ unsigned int pkt_len)
+{
+ struct nfp_net_tx_buf *txbuf;
+ struct nfp_net_tx_desc *txd;
+ dma_addr_t new_dma_addr;
+ void *new_frag;
+ int wr_idx;
+
+ if (unlikely(nfp_net_tx_full(tx_ring, 1))) {
+ nfp_net_rx_drop(rx_ring->r_vec, rx_ring, rxbuf, NULL);
+ return;
+ }
+
+ new_frag = nfp_net_napi_alloc_one(nn, DMA_BIDIRECTIONAL, &new_dma_addr);
+ if (unlikely(!new_frag)) {
+ nfp_net_rx_drop(rx_ring->r_vec, rx_ring, rxbuf, NULL);
+ return;
+ }
+ nfp_net_rx_give_one(rx_ring, new_frag, new_dma_addr);
+
+ wr_idx = tx_ring->wr_p & (tx_ring->cnt - 1);
+
+ /* Stash the soft descriptor of the head then initialize it */
+ txbuf = &tx_ring->txbufs[wr_idx];
+ txbuf->frag = rxbuf->frag;
+ txbuf->dma_addr = rxbuf->dma_addr;
+ txbuf->fidx = -1;
+ txbuf->pkt_cnt = 1;
+ txbuf->real_len = pkt_len;
+
+ dma_sync_single_for_device(&nn->pdev->dev, rxbuf->dma_addr + pkt_off,
+ pkt_len, DMA_TO_DEVICE);
+
+ /* Build TX descriptor */
+ txd = &tx_ring->txds[wr_idx];
+ txd->offset_eop = PCIE_DESC_TX_EOP;
+ txd->dma_len = cpu_to_le16(pkt_len);
+ nfp_desc_set_dma_addr(txd, rxbuf->dma_addr + pkt_off);
+ txd->data_len = cpu_to_le16(pkt_len);
+
+ txd->flags = 0;
+ txd->mss = 0;
+ txd->l4_offset = 0;
+
+ tx_ring->wr_p++;
+ tx_ring->wr_ptr_add++;
+}
+
+static int nfp_net_run_xdp(struct bpf_prog *prog, void *data, unsigned int len)
+{
+ struct xdp_buff xdp;
+
+ xdp.data = data;
+ xdp.data_end = data + len;
+
+ return BPF_PROG_RUN(prog, (void *)&xdp);
+}
+
/**
* nfp_net_rx() - receive up to @budget packets on @rx_ring
* @rx_ring: RX ring to receive from
@@ -1383,10 +1536,20 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
{
struct nfp_net_r_vector *r_vec = rx_ring->r_vec;
struct nfp_net *nn = r_vec->nfp_net;
+ struct nfp_net_tx_ring *tx_ring;
+ struct bpf_prog *xdp_prog;
+ unsigned int true_bufsz;
struct sk_buff *skb;
int pkts_polled = 0;
+ int rx_dma_map_dir;
int idx;
+ rcu_read_lock();
+ xdp_prog = READ_ONCE(nn->xdp_prog);
+ rx_dma_map_dir = xdp_prog ? DMA_BIDIRECTIONAL : DMA_FROM_DEVICE;
+ true_bufsz = xdp_prog ? PAGE_SIZE : nn->fl_bufsz;
+ tx_ring = r_vec->xdp_ring;
+
while (pkts_polled < budget) {
unsigned int meta_len, data_len, data_off, pkt_len, pkt_off;
struct nfp_net_rx_buf *rxbuf;
@@ -1437,19 +1600,45 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
r_vec->rx_bytes += pkt_len;
u64_stats_update_end(&r_vec->rx_sync);
- skb = build_skb(rxbuf->frag, nn->fl_bufsz);
+ if (xdp_prog) {
+ int act;
+
+ dma_sync_single_for_cpu(&nn->pdev->dev,
+ rxbuf->dma_addr + pkt_off,
+ pkt_len, DMA_FROM_DEVICE);
+ act = nfp_net_run_xdp(xdp_prog, rxbuf->frag + data_off,
+ pkt_len);
+ switch (act) {
+ case XDP_PASS:
+ break;
+ case XDP_TX:
+ nfp_net_tx_xdp_buf(nn, rx_ring, tx_ring, rxbuf,
+ pkt_off, pkt_len);
+ continue;
+ default:
+ bpf_warn_invalid_xdp_action(act);
+ case XDP_ABORTED:
+ case XDP_DROP:
+ nfp_net_rx_give_one(rx_ring, rxbuf->frag,
+ rxbuf->dma_addr);
+ continue;
+ }
+ }
+
+ skb = build_skb(rxbuf->frag, true_bufsz);
if (unlikely(!skb)) {
nfp_net_rx_drop(r_vec, rx_ring, rxbuf, NULL);
continue;
}
- new_frag = nfp_net_napi_alloc_one(nn, &new_dma_addr);
+ new_frag = nfp_net_napi_alloc_one(nn, rx_dma_map_dir,
+ &new_dma_addr);
if (unlikely(!new_frag)) {
nfp_net_rx_drop(r_vec, rx_ring, rxbuf, skb);
continue;
}
- nfp_net_dma_unmap_rx(nn, rx_ring->rxbufs[idx].dma_addr,
- nn->fl_bufsz, DMA_FROM_DEVICE);
+ nfp_net_dma_unmap_rx(nn, rxbuf->dma_addr, nn->fl_bufsz,
+ rx_dma_map_dir);
nfp_net_rx_give_one(rx_ring, new_frag, new_dma_addr);
@@ -1481,6 +1670,10 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
napi_gro_receive(&rx_ring->r_vec->napi, skb);
}
+ if (xdp_prog && tx_ring->wr_ptr_add)
+ nfp_net_tx_xmit_more_flush(tx_ring);
+ rcu_read_unlock();
+
return pkts_polled;
}
@@ -1499,8 +1692,11 @@ static int nfp_net_poll(struct napi_struct *napi, int budget)
if (r_vec->tx_ring)
nfp_net_tx_complete(r_vec->tx_ring);
- if (r_vec->rx_ring)
+ if (r_vec->rx_ring) {
pkts_polled = nfp_net_rx(r_vec->rx_ring, budget);
+ if (r_vec->xdp_ring)
+ nfp_net_xdp_complete(r_vec->xdp_ring);
+ }
if (pkts_polled < budget) {
napi_complete_done(napi, pkts_polled);
@@ -1540,10 +1736,12 @@ static void nfp_net_tx_ring_free(struct nfp_net_tx_ring *tx_ring)
* nfp_net_tx_ring_alloc() - Allocate resource for a TX ring
* @tx_ring: TX Ring structure to allocate
* @cnt: Ring buffer count
+ * @is_xdp: True if ring will be used for XDP
*
* Return: 0 on success, negative errno otherwise.
*/
-static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
+static int
+nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt, bool is_xdp)
{
struct nfp_net_r_vector *r_vec = tx_ring->r_vec;
struct nfp_net *nn = r_vec->nfp_net;
@@ -1563,11 +1761,14 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
if (!tx_ring->txbufs)
goto err_alloc;
- netif_set_xps_queue(nn->netdev, &r_vec->affinity_mask, tx_ring->idx);
+ if (!is_xdp)
+ netif_set_xps_queue(nn->netdev, &r_vec->affinity_mask,
+ tx_ring->idx);
- nn_dbg(nn, "TxQ%02d: QCidx=%02d cnt=%d dma=%#llx host=%p\n",
+ nn_dbg(nn, "TxQ%02d: QCidx=%02d cnt=%d dma=%#llx host=%p %s\n",
tx_ring->idx, tx_ring->qcidx,
- tx_ring->cnt, (unsigned long long)tx_ring->dma, tx_ring->txds);
+ tx_ring->cnt, (unsigned long long)tx_ring->dma, tx_ring->txds,
+ is_xdp ? "XDP" : "");
return 0;
@@ -1577,7 +1778,8 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
}
static struct nfp_net_tx_ring *
-nfp_net_tx_ring_set_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_tx_ring_set_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s,
+ unsigned int num_stack_tx_rings)
{
struct nfp_net_tx_ring *rings;
unsigned int r;
@@ -1587,9 +1789,14 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
return NULL;
for (r = 0; r < s->n_rings; r++) {
- nfp_net_tx_ring_init(&rings[r], &nn->r_vecs[r], r);
+ int bias = 0;
+
+ if (r >= num_stack_tx_rings)
+ bias = num_stack_tx_rings;
- if (nfp_net_tx_ring_alloc(&rings[r], s->dcnt))
+ nfp_net_tx_ring_init(&rings[r], &nn->r_vecs[r - bias], r);
+
+ if (nfp_net_tx_ring_alloc(&rings[r], s->dcnt, bias))
goto err_free_prev;
}
@@ -1694,7 +1901,8 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
}
static struct nfp_net_rx_ring *
-nfp_net_rx_ring_set_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_rx_ring_set_prepare(struct nfp_net *nn, struct nfp_net_ring_set *s,
+ bool xdp)
{
unsigned int fl_bufsz = nfp_net_calc_fl_bufsz(nn, s->mtu);
struct nfp_net_rx_ring *rings;
@@ -1710,7 +1918,7 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
if (nfp_net_rx_ring_alloc(&rings[r], fl_bufsz, s->dcnt))
goto err_free_prev;
- if (nfp_net_rx_ring_bufs_alloc(nn, &rings[r]))
+ if (nfp_net_rx_ring_bufs_alloc(nn, &rings[r], xdp))
goto err_free_ring;
}
@@ -1718,7 +1926,7 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
err_free_prev:
while (r--) {
- nfp_net_rx_ring_bufs_free(nn, &rings[r]);
+ nfp_net_rx_ring_bufs_free(nn, &rings[r], xdp);
err_free_ring:
nfp_net_rx_ring_free(&rings[r]);
}
@@ -1744,13 +1952,14 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
}
static void
-nfp_net_rx_ring_set_free(struct nfp_net *nn, struct nfp_net_ring_set *s)
+nfp_net_rx_ring_set_free(struct nfp_net *nn, struct nfp_net_ring_set *s,
+ bool xdp)
{
struct nfp_net_rx_ring *rings = s->rings;
unsigned int r;
for (r = 0; r < s->n_rings; r++) {
- nfp_net_rx_ring_bufs_free(nn, &rings[r]);
+ nfp_net_rx_ring_bufs_free(nn, &rings[r], xdp);
nfp_net_rx_ring_free(&rings[r]);
}
@@ -1762,7 +1971,11 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
int idx)
{
r_vec->rx_ring = idx < nn->num_rx_rings ? &nn->rx_rings[idx] : NULL;
- r_vec->tx_ring = idx < nn->num_tx_rings ? &nn->tx_rings[idx] : NULL;
+ r_vec->tx_ring =
+ idx < nn->num_stack_tx_rings ? &nn->tx_rings[idx] : NULL;
+
+ r_vec->xdp_ring = idx < nn->num_tx_rings - nn->num_stack_tx_rings ?
+ &nn->tx_rings[nn->num_stack_tx_rings + idx] : NULL;
}
static int
@@ -2083,13 +2296,14 @@ static int nfp_net_netdev_open(struct net_device *netdev)
goto err_cleanup_vec_p;
}
- nn->rx_rings = nfp_net_rx_ring_set_prepare(nn, &rx);
+ nn->rx_rings = nfp_net_rx_ring_set_prepare(nn, &rx, nn->xdp_prog);
if (!nn->rx_rings) {
err = -ENOMEM;
goto err_cleanup_vec;
}
- nn->tx_rings = nfp_net_tx_ring_set_prepare(nn, &tx);
+ nn->tx_rings = nfp_net_tx_ring_set_prepare(nn, &tx,
+ nn->num_stack_tx_rings);
if (!nn->tx_rings) {
err = -ENOMEM;
goto err_free_rx_rings;
@@ -2098,7 +2312,7 @@ static int nfp_net_netdev_open(struct net_device *netdev)
for (r = 0; r < nn->max_r_vecs; r++)
nfp_net_vector_assign_rings(nn, &nn->r_vecs[r], r);
- err = netif_set_real_num_tx_queues(netdev, nn->num_tx_rings);
+ err = netif_set_real_num_tx_queues(netdev, nn->num_stack_tx_rings);
if (err)
goto err_free_rings;
@@ -2130,7 +2344,7 @@ static int nfp_net_netdev_open(struct net_device *netdev)
err_free_rings:
nfp_net_tx_ring_set_free(nn, &tx);
err_free_rx_rings:
- nfp_net_rx_ring_set_free(nn, &rx);
+ nfp_net_rx_ring_set_free(nn, &rx, nn->xdp_prog);
err_cleanup_vec:
r = nn->num_r_vecs;
err_cleanup_vec_p:
@@ -2171,7 +2385,7 @@ static void nfp_net_close_free_all(struct nfp_net *nn)
unsigned int r;
for (r = 0; r < nn->num_rx_rings; r++) {
- nfp_net_rx_ring_bufs_free(nn, &nn->rx_rings[r]);
+ nfp_net_rx_ring_bufs_free(nn, &nn->rx_rings[r], nn->xdp_prog);
nfp_net_rx_ring_free(&nn->rx_rings[r]);
}
for (r = 0; r < nn->num_tx_rings; r++)
@@ -2251,6 +2465,8 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
static int
nfp_net_ring_swap_enable(struct nfp_net *nn, unsigned int *num_vecs,
+ unsigned int *stack_tx_rings,
+ struct bpf_prog **xdp_prog,
struct nfp_net_ring_set *rx,
struct nfp_net_ring_set *tx)
{
@@ -2263,6 +2479,8 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
nfp_net_tx_ring_set_swap(nn, tx);
swap(*num_vecs, nn->num_r_vecs);
+ swap(*stack_tx_rings, nn->num_stack_tx_rings);
+ *xdp_prog = xchg(&nn->xdp_prog, *xdp_prog);
for (r = 0; r < nn->max_r_vecs; r++)
nfp_net_vector_assign_rings(nn, &nn->r_vecs[r], r);
@@ -2277,9 +2495,9 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
return err;
}
- if (nn->netdev->real_num_tx_queues != nn->num_tx_rings) {
+ if (nn->netdev->real_num_tx_queues != nn->num_stack_tx_rings) {
err = netif_set_real_num_tx_queues(nn->netdev,
- nn->num_tx_rings);
+ nn->num_stack_tx_rings);
if (err)
return err;
}
@@ -2287,11 +2505,30 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
return __nfp_net_set_config_and_enable(nn);
}
+static int
+nfp_net_check_config(struct nfp_net *nn, struct bpf_prog *xdp_prog,
+ struct nfp_net_ring_set *rx, struct nfp_net_ring_set *tx)
+{
+ /* XDP-enabled tests */
+ if (!xdp_prog)
+ return 0;
+ if (rx && nfp_net_calc_fl_bufsz(nn, rx->mtu) > PAGE_SIZE) {
+ nn_warn(nn, "MTU too large w/ XDP enabled\n");
+ return -EINVAL;
+ }
+ if (tx && tx->n_rings > nn->max_tx_rings) {
+ nn_warn(nn, "Insufficient number of TX rings w/ XDP enabled\n");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
static void
-nfp_net_ring_reconfig_down(struct nfp_net *nn,
+nfp_net_ring_reconfig_down(struct nfp_net *nn, struct bpf_prog **xdp_prog,
struct nfp_net_ring_set *rx,
struct nfp_net_ring_set *tx,
- unsigned int num_vecs)
+ unsigned int stack_tx_rings, unsigned int num_vecs)
{
nn->netdev->mtu = rx ? rx->mtu : nn->netdev->mtu;
nn->fl_bufsz = nfp_net_calc_fl_bufsz(nn, nn->netdev->mtu);
@@ -2299,24 +2536,34 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
nn->txd_cnt = tx ? tx->dcnt : nn->txd_cnt;
nn->num_rx_rings = rx ? rx->n_rings : nn->num_rx_rings;
nn->num_tx_rings = tx ? tx->n_rings : nn->num_tx_rings;
+ nn->num_stack_tx_rings = stack_tx_rings;
nn->num_r_vecs = num_vecs;
+ *xdp_prog = xchg(&nn->xdp_prog, *xdp_prog);
if (!netif_is_rxfh_configured(nn->netdev))
nfp_net_rss_init_itbl(nn);
}
int
-nfp_net_ring_reconfig(struct nfp_net *nn, struct nfp_net_ring_set *rx,
- struct nfp_net_ring_set *tx)
+nfp_net_ring_reconfig(struct nfp_net *nn, struct bpf_prog **xdp_prog,
+ struct nfp_net_ring_set *rx, struct nfp_net_ring_set *tx)
{
- unsigned int num_vecs, r;
+ unsigned int stack_tx_rings, num_vecs, r;
int err;
- num_vecs = max(rx ? rx->n_rings : nn->num_rx_rings,
- tx ? tx->n_rings : nn->num_tx_rings);
+ stack_tx_rings = tx ? tx->n_rings : nn->num_tx_rings;
+ if (*xdp_prog)
+ stack_tx_rings -= rx ? rx->n_rings : nn->num_rx_rings;
+
+ num_vecs = max(rx ? rx->n_rings : nn->num_rx_rings, stack_tx_rings);
+
+ err = nfp_net_check_config(nn, *xdp_prog, rx, tx);
+ if (err)
+ return err;
if (!netif_running(nn->netdev)) {
- nfp_net_ring_reconfig_down(nn, rx, tx, num_vecs);
+ nfp_net_ring_reconfig_down(nn, xdp_prog, rx, tx,
+ stack_tx_rings, num_vecs);
return 0;
}
@@ -2329,13 +2576,13 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
}
}
if (rx) {
- if (!nfp_net_rx_ring_set_prepare(nn, rx)) {
+ if (!nfp_net_rx_ring_set_prepare(nn, rx, *xdp_prog)) {
err = -ENOMEM;
goto err_cleanup_vecs;
}
}
if (tx) {
- if (!nfp_net_tx_ring_set_prepare(nn, tx)) {
+ if (!nfp_net_tx_ring_set_prepare(nn, tx, stack_tx_rings)) {
err = -ENOMEM;
goto err_free_rx;
}
@@ -2345,14 +2592,16 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
nfp_net_close_stack(nn);
nfp_net_clear_config_and_disable(nn);
- err = nfp_net_ring_swap_enable(nn, &num_vecs, rx, tx);
+ err = nfp_net_ring_swap_enable(nn, &num_vecs, &stack_tx_rings,
+ xdp_prog, rx, tx);
if (err) {
int err2;
nfp_net_clear_config_and_disable(nn);
/* Try with old configuration and old rings */
- err2 = nfp_net_ring_swap_enable(nn, &num_vecs, rx, tx);
+ err2 = nfp_net_ring_swap_enable(nn, &num_vecs, &stack_tx_rings,
+ xdp_prog, rx, tx);
if (err2)
nn_err(nn, "Can't restore ring config - FW communication failed (%d,%d)\n",
err, err2);
@@ -2361,7 +2610,7 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
nfp_net_cleanup_vector(nn, &nn->r_vecs[r]);
if (rx)
- nfp_net_rx_ring_set_free(nn, rx);
+ nfp_net_rx_ring_set_free(nn, rx, *xdp_prog);
if (tx)
nfp_net_tx_ring_set_free(nn, tx);
@@ -2371,7 +2620,7 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
err_free_rx:
if (rx)
- nfp_net_rx_ring_set_free(nn, rx);
+ nfp_net_rx_ring_set_free(nn, rx, *xdp_prog);
err_cleanup_vecs:
for (r = num_vecs - 1; r >= nn->num_r_vecs; r--)
nfp_net_cleanup_vector(nn, &nn->r_vecs[r]);
@@ -2387,7 +2636,7 @@ static int nfp_net_change_mtu(struct net_device *netdev, int new_mtu)
.dcnt = nn->rxd_cnt,
};
- return nfp_net_ring_reconfig(nn, &rx, NULL);
+ return nfp_net_ring_reconfig(nn, &nn->xdp_prog, &rx, NULL);
}
static struct rtnl_link_stats64 *nfp_net_stat64(struct net_device *netdev,
@@ -2653,6 +2902,56 @@ static void nfp_net_del_vxlan_port(struct net_device *netdev,
nfp_net_set_vxlan_port(nn, idx, 0);
}
+static int nfp_net_xdp_setup(struct nfp_net *nn, struct bpf_prog *prog)
+{
+ struct nfp_net_ring_set rx = {
+ .n_rings = nn->num_rx_rings,
+ .mtu = nn->netdev->mtu,
+ .dcnt = nn->rxd_cnt,
+ };
+ struct nfp_net_ring_set tx = {
+ .n_rings = nn->num_tx_rings,
+ .dcnt = nn->txd_cnt,
+ };
+ int err;
+
+ if (!prog && !nn->xdp_prog)
+ return 0;
+ if (prog && nn->xdp_prog) {
+ prog = xchg(&nn->xdp_prog, prog);
+ bpf_prog_put(prog);
+ return 0;
+ }
+
+ tx.n_rings += prog ? nn->num_rx_rings : -nn->num_rx_rings;
+
+ /* We need RX reconfig to remap the buffers (BIDIR vs FROM_DEV) */
+ err = nfp_net_ring_reconfig(nn, &prog, &rx, &tx);
+ if (err)
+ return err;
+
+ /* @prog got swapped and is now the old one */
+ if (prog)
+ bpf_prog_put(prog);
+
+ return 0;
+}
+
+static int nfp_net_xdp(struct net_device *netdev, struct netdev_xdp *xdp)
+{
+ struct nfp_net *nn = netdev_priv(netdev);
+
+ switch (xdp->command) {
+ case XDP_SETUP_PROG:
+ return nfp_net_xdp_setup(nn, xdp->prog);
+ case XDP_QUERY_PROG:
+ xdp->prog_attached = !!nn->xdp_prog;
+ return 0;
+ default:
+ return -EINVAL;
+ }
+}
+
static const struct net_device_ops nfp_net_netdev_ops = {
.ndo_open = nfp_net_netdev_open,
.ndo_stop = nfp_net_netdev_close,
@@ -2667,6 +2966,7 @@ static void nfp_net_del_vxlan_port(struct net_device *netdev,
.ndo_features_check = nfp_net_features_check,
.ndo_udp_tunnel_add = nfp_net_add_vxlan_port,
.ndo_udp_tunnel_del = nfp_net_del_vxlan_port,
+ .ndo_xdp = nfp_net_xdp,
};
/**
@@ -2929,5 +3229,9 @@ int nfp_net_netdev_init(struct net_device *netdev)
*/
void nfp_net_netdev_clean(struct net_device *netdev)
{
- unregister_netdev(netdev);
+ struct nfp_net *nn = netdev_priv(netdev);
+
+ if (nn->xdp_prog)
+ bpf_prog_put(nn->xdp_prog);
+ unregister_netdev(nn->netdev);
}
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
index e90df8145323..c66f3f954aa8 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
@@ -114,6 +114,16 @@ static int nfp_net_debugfs_rx_q_open(struct inode *inode, struct file *f)
.llseek = seq_lseek
};
+static int nfp_net_debugfs_tx_q_open(struct inode *inode, struct file *f);
+
+static const struct file_operations nfp_tx_q_fops = {
+ .owner = THIS_MODULE,
+ .open = nfp_net_debugfs_tx_q_open,
+ .release = single_release,
+ .read = seq_read,
+ .llseek = seq_lseek
+};
+
static int nfp_net_debugfs_tx_q_read(struct seq_file *file, void *data)
{
struct nfp_net_r_vector *r_vec = file->private;
@@ -126,10 +136,13 @@ static int nfp_net_debugfs_tx_q_read(struct seq_file *file, void *data)
rtnl_lock();
- if (!r_vec->nfp_net || !r_vec->tx_ring)
+ if (debugfs_real_fops(file->file) == &nfp_tx_q_fops)
+ tx_ring = r_vec->tx_ring;
+ else
+ tx_ring = r_vec->xdp_ring;
+ if (!r_vec->nfp_net || !tx_ring)
goto out;
nn = r_vec->nfp_net;
- tx_ring = r_vec->tx_ring;
if (!netif_running(nn->netdev))
goto out;
@@ -148,9 +161,14 @@ static int nfp_net_debugfs_tx_q_read(struct seq_file *file, void *data)
txd->vals[2], txd->vals[3]);
skb = READ_ONCE(tx_ring->txbufs[i].skb);
- if (skb)
- seq_printf(file, " skb->head=%p skb->data=%p",
- skb->head, skb->data);
+ if (skb) {
+ if (tx_ring == r_vec->tx_ring)
+ seq_printf(file, " skb->head=%p skb->data=%p",
+ skb->head, skb->data);
+ else
+ seq_printf(file, " frag=%p", skb);
+ }
+
if (tx_ring->txbufs[i].dma_addr)
seq_printf(file, " dma_addr=%pad",
&tx_ring->txbufs[i].dma_addr);
@@ -176,7 +194,7 @@ static int nfp_net_debugfs_tx_q_open(struct inode *inode, struct file *f)
return single_open(f, nfp_net_debugfs_tx_q_read, inode->i_private);
}
-static const struct file_operations nfp_tx_q_fops = {
+static const struct file_operations nfp_xdp_q_fops = {
.owner = THIS_MODULE,
.open = nfp_net_debugfs_tx_q_open,
.release = single_release,
@@ -186,7 +204,7 @@ static int nfp_net_debugfs_tx_q_open(struct inode *inode, struct file *f)
void nfp_net_debugfs_adapter_add(struct nfp_net *nn)
{
- struct dentry *queues, *tx, *rx;
+ struct dentry *queues, *tx, *rx, *xdp;
char int_name[16];
int i;
@@ -204,13 +222,16 @@ void nfp_net_debugfs_adapter_add(struct nfp_net *nn)
rx = debugfs_create_dir("rx", queues);
tx = debugfs_create_dir("tx", queues);
- if (IS_ERR_OR_NULL(rx) || IS_ERR_OR_NULL(tx))
+ xdp = debugfs_create_dir("xdp", queues);
+ if (IS_ERR_OR_NULL(rx) || IS_ERR_OR_NULL(tx) || IS_ERR_OR_NULL(xdp))
return;
for (i = 0; i < min(nn->max_rx_rings, nn->max_r_vecs); i++) {
sprintf(int_name, "%d", i);
debugfs_create_file(int_name, S_IRUSR, rx,
&nn->r_vecs[i], &nfp_rx_q_fops);
+ debugfs_create_file(int_name, S_IRUSR, xdp,
+ &nn->r_vecs[i], &nfp_xdp_q_fops);
}
for (i = 0; i < min(nn->max_tx_rings, nn->max_r_vecs); i++) {
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
index b87f1b73f200..1b26e9646574 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
@@ -176,7 +176,8 @@ static int nfp_net_set_ring_size(struct nfp_net *nn, u32 rxd_cnt, u32 txd_cnt)
if (nn->txd_cnt != txd_cnt)
reconfig_tx = &tx;
- return nfp_net_ring_reconfig(nn, reconfig_rx, reconfig_tx);
+ return nfp_net_ring_reconfig(nn, &nn->xdp_prog,
+ reconfig_rx, reconfig_tx);
}
static int nfp_net_set_ringparam(struct net_device *netdev,
@@ -639,14 +640,19 @@ static void nfp_net_get_channels(struct net_device *netdev,
struct ethtool_channels *channel)
{
struct nfp_net *nn = netdev_priv(netdev);
+ unsigned int num_tx_rings;
+
+ num_tx_rings = nn->num_tx_rings;
+ if (nn->xdp_prog)
+ num_tx_rings -= nn->num_rx_rings;
channel->max_rx = min(nn->max_rx_rings, nn->max_r_vecs);
channel->max_tx = min(nn->max_tx_rings, nn->max_r_vecs);
channel->max_combined = min(channel->max_rx, channel->max_tx);
channel->max_other = NFP_NET_NON_Q_VECTORS;
- channel->combined_count = min(nn->num_rx_rings, nn->num_tx_rings);
+ channel->combined_count = min(nn->num_rx_rings, num_tx_rings);
channel->rx_count = nn->num_rx_rings - channel->combined_count;
- channel->tx_count = nn->num_tx_rings - channel->combined_count;
+ channel->tx_count = num_tx_rings - channel->combined_count;
channel->other_count = NFP_NET_NON_Q_VECTORS;
}
@@ -666,10 +672,16 @@ static int nfp_net_set_num_rings(struct nfp_net *nn, unsigned int total_rx,
if (nn->num_rx_rings != total_rx)
reconfig_rx = ℞
- if (nn->num_tx_rings != total_tx)
+ if (nn->num_stack_tx_rings != total_tx ||
+ (nn->xdp_prog && reconfig_rx))
reconfig_tx = &tx;
- return nfp_net_ring_reconfig(nn, reconfig_rx, reconfig_tx);
+ /* nfp_net_check_config() will catch tx.n_rings > nn->max_tx_rings */
+ if (nn->xdp_prog)
+ tx.n_rings += total_rx;
+
+ return nfp_net_ring_reconfig(nn, &nn->xdp_prog,
+ reconfig_rx, reconfig_tx);
}
static int nfp_net_set_channels(struct net_device *netdev,
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 09/13] nfp: reorganize nfp_net_rx() to get packet offsets early
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
Calculate packet offsets early in nfp_net_rx() so that we will be
able to use them in upcoming XDP handler. While at it move relevant
variables into the loop scope.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
.../net/ethernet/netronome/nfp/nfp_net_common.c | 56 ++++++++++++----------
1 file changed, 30 insertions(+), 26 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 506362729607..2ab63661a6fd 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -1383,16 +1383,17 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
{
struct nfp_net_r_vector *r_vec = rx_ring->r_vec;
struct nfp_net *nn = r_vec->nfp_net;
- unsigned int data_len, meta_len;
- struct nfp_net_rx_buf *rxbuf;
- struct nfp_net_rx_desc *rxd;
- dma_addr_t new_dma_addr;
struct sk_buff *skb;
int pkts_polled = 0;
- void *new_frag;
int idx;
while (pkts_polled < budget) {
+ unsigned int meta_len, data_len, data_off, pkt_len, pkt_off;
+ struct nfp_net_rx_buf *rxbuf;
+ struct nfp_net_rx_desc *rxd;
+ dma_addr_t new_dma_addr;
+ void *new_frag;
+
idx = rx_ring->rd_p & (rx_ring->cnt - 1);
rxd = &rx_ring->rxds[idx];
@@ -1408,22 +1409,6 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
pkts_polled++;
rxbuf = &rx_ring->rxbufs[idx];
- skb = build_skb(rxbuf->frag, nn->fl_bufsz);
- if (unlikely(!skb)) {
- nfp_net_rx_drop(r_vec, rx_ring, rxbuf, NULL);
- continue;
- }
- new_frag = nfp_net_napi_alloc_one(nn, &new_dma_addr);
- if (unlikely(!new_frag)) {
- nfp_net_rx_drop(r_vec, rx_ring, rxbuf, skb);
- continue;
- }
-
- nfp_net_dma_unmap_rx(nn, rx_ring->rxbufs[idx].dma_addr,
- nn->fl_bufsz, DMA_FROM_DEVICE);
-
- nfp_net_rx_give_one(rx_ring, new_frag, new_dma_addr);
-
/* < meta_len >
* <-- [rx_offset] -->
* ---------------------------------------------------------
@@ -1438,20 +1423,39 @@ static int nfp_net_rx(struct nfp_net_rx_ring *rx_ring, int budget)
*/
meta_len = rxd->rxd.meta_len_dd & PCIE_DESC_RX_META_LEN_MASK;
data_len = le16_to_cpu(rxd->rxd.data_len);
+ pkt_len = data_len - meta_len;
if (nn->rx_offset == NFP_NET_CFG_RX_OFFSET_DYNAMIC)
- skb_reserve(skb, NFP_NET_RX_BUF_HEADROOM + meta_len);
+ pkt_off = meta_len;
else
- skb_reserve(skb,
- NFP_NET_RX_BUF_HEADROOM + nn->rx_offset);
- skb_put(skb, data_len - meta_len);
+ pkt_off = nn->rx_offset;
+ data_off = NFP_NET_RX_BUF_HEADROOM + pkt_off;
/* Stats update */
u64_stats_update_begin(&r_vec->rx_sync);
r_vec->rx_pkts++;
- r_vec->rx_bytes += skb->len;
+ r_vec->rx_bytes += pkt_len;
u64_stats_update_end(&r_vec->rx_sync);
+ skb = build_skb(rxbuf->frag, nn->fl_bufsz);
+ if (unlikely(!skb)) {
+ nfp_net_rx_drop(r_vec, rx_ring, rxbuf, NULL);
+ continue;
+ }
+ new_frag = nfp_net_napi_alloc_one(nn, &new_dma_addr);
+ if (unlikely(!new_frag)) {
+ nfp_net_rx_drop(r_vec, rx_ring, rxbuf, skb);
+ continue;
+ }
+
+ nfp_net_dma_unmap_rx(nn, rx_ring->rxbufs[idx].dma_addr,
+ nn->fl_bufsz, DMA_FROM_DEVICE);
+
+ nfp_net_rx_give_one(rx_ring, new_frag, new_dma_addr);
+
+ skb_reserve(skb, data_off);
+ skb_put(skb, pkt_len);
+
if (nn->fw_ver.major <= 3) {
nfp_net_set_hash_desc(nn->netdev, skb, rxd);
} else if (meta_len) {
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 12/13] nfp: remove unnecessary parameters from nfp_net_bpf_offload()
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
nfp_net_bpf_offload() takes all .setup_tc() parameters but it
doesn't use them at the moment. Remove unnecessary ones to make
it possible for XDP to reuse this function.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/nfp_net.h | 4 +---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 2 +-
drivers/net/ethernet/netronome/nfp/nfp_net_offload.c | 4 +---
3 files changed, 3 insertions(+), 7 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
index abc9e56e93b8..fd29a6306991 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
@@ -817,8 +817,6 @@ static inline void nfp_net_debugfs_adapter_del(struct nfp_net *nn)
#endif /* CONFIG_NFP_NET_DEBUG */
void nfp_net_filter_stats_timer(unsigned long data);
-int
-nfp_net_bpf_offload(struct nfp_net *nn, u32 handle, __be16 proto,
- struct tc_cls_bpf_offload *cls_bpf);
+int nfp_net_bpf_offload(struct nfp_net *nn, struct tc_cls_bpf_offload *cls_bpf);
#endif /* _NFP_NET_H_ */
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index fa43dbcecc4f..1e8e00d25c51 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -2694,7 +2694,7 @@ static bool nfp_net_ebpf_capable(struct nfp_net *nn)
return -ENOTSUPP;
if (tc->type == TC_SETUP_CLSBPF && nfp_net_ebpf_capable(nn))
- return nfp_net_bpf_offload(nn, handle, proto, tc->cls_bpf);
+ return nfp_net_bpf_offload(nn, tc->cls_bpf);
return -EINVAL;
}
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c b/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c
index cfed40c0e310..4bb6f16e2a7a 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_offload.c
@@ -233,9 +233,7 @@ static int nfp_net_bpf_stop(struct nfp_net *nn)
return nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_GEN);
}
-int
-nfp_net_bpf_offload(struct nfp_net *nn, u32 handle, __be16 proto,
- struct tc_cls_bpf_offload *cls_bpf)
+int nfp_net_bpf_offload(struct nfp_net *nn, struct tc_cls_bpf_offload *cls_bpf)
{
struct nfp_bpf_result res;
dma_addr_t dma_addr;
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 10/13] debugfs: constify argument to debugfs_real_fops()
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev
Cc: Jakub Kicinski, Greg Kroah-Hartman, Nicolai Stange,
Christian Lamparter, LKML
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
seq_file users can only access const version of file pointer,
because the ->file member of struct seq_operations is marked
as such. Make parameter to debugfs_real_fops() const.
CC: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CC: Nicolai Stange <nicstange@gmail.com>
CC: Christian Lamparter <chunkeey@gmail.com>
CC: LKML <linux-kernel@vger.kernel.org>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
include/linux/debugfs.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/include/linux/debugfs.h b/include/linux/debugfs.h
index 4d3f0d1aec73..bf1907d96097 100644
--- a/include/linux/debugfs.h
+++ b/include/linux/debugfs.h
@@ -52,7 +52,8 @@ struct debugfs_regset32 {
* Must only be called under the protection established by
* debugfs_use_file_start().
*/
-static inline const struct file_operations *debugfs_real_fops(struct file *filp)
+static inline const struct file_operations *
+debugfs_real_fops(const struct file *filp)
__must_hold(&debugfs_srcu)
{
/*
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 08/13] nfp: add support for ethtool .set_channels
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
Allow changing the number of rings via ethtool .set_channels API.
Runtime reconfig needs to be extended to handle number of rings.
We need to be able to activate interrupt vectors before rings are
assigned to them.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/nfp_net.h | 1 +
.../net/ethernet/netronome/nfp/nfp_net_common.c | 93 +++++++++++++++++-----
.../net/ethernet/netronome/nfp/nfp_net_debugfs.c | 4 +-
.../net/ethernet/netronome/nfp/nfp_net_ethtool.c | 47 +++++++++++
4 files changed, 123 insertions(+), 22 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net.h b/drivers/net/ethernet/netronome/nfp/nfp_net.h
index 14b5e21cabf1..486e7c6453bc 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net.h
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net.h
@@ -584,6 +584,7 @@ struct nfp_net {
};
struct nfp_net_ring_set {
+ unsigned int n_rings;
unsigned int mtu;
unsigned int dcnt;
void *rings;
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 09cec6e2c6cf..506362729607 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -494,7 +494,7 @@ static void nfp_net_irqs_assign(struct net_device *netdev)
nn->lsc_handler = nfp_net_irq_lsc;
nn->exn_handler = nfp_net_irq_exn;
- for (r = 0; r < nn->num_r_vecs; r++) {
+ for (r = 0; r < nn->max_r_vecs; r++) {
r_vec = &nn->r_vecs[r];
r_vec->nfp_net = nn;
r_vec->handler = nfp_net_irq_rxtx;
@@ -1578,12 +1578,12 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
struct nfp_net_tx_ring *rings;
unsigned int r;
- rings = kcalloc(nn->num_tx_rings, sizeof(*rings), GFP_KERNEL);
+ rings = kcalloc(s->n_rings, sizeof(*rings), GFP_KERNEL);
if (!rings)
return NULL;
- for (r = 0; r < nn->num_tx_rings; r++) {
- nfp_net_tx_ring_init(&rings[r], nn->tx_rings[r].r_vec, r);
+ for (r = 0; r < s->n_rings; r++) {
+ nfp_net_tx_ring_init(&rings[r], &nn->r_vecs[r], r);
if (nfp_net_tx_ring_alloc(&rings[r], s->dcnt))
goto err_free_prev;
@@ -1605,9 +1605,11 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
s->dcnt = nn->txd_cnt;
s->rings = nn->tx_rings;
+ s->n_rings = nn->num_tx_rings;
nn->txd_cnt = new.dcnt;
nn->tx_rings = new.rings;
+ nn->num_tx_rings = new.n_rings;
}
static void
@@ -1616,7 +1618,7 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
struct nfp_net_tx_ring *rings = s->rings;
unsigned int r;
- for (r = 0; r < nn->num_tx_rings; r++)
+ for (r = 0; r < s->n_rings; r++)
nfp_net_tx_ring_free(&rings[r]);
kfree(rings);
@@ -1694,12 +1696,12 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
struct nfp_net_rx_ring *rings;
unsigned int r;
- rings = kcalloc(nn->num_rx_rings, sizeof(*rings), GFP_KERNEL);
+ rings = kcalloc(s->n_rings, sizeof(*rings), GFP_KERNEL);
if (!rings)
return NULL;
- for (r = 0; r < nn->num_rx_rings; r++) {
- nfp_net_rx_ring_init(&rings[r], nn->rx_rings[r].r_vec, r);
+ for (r = 0; r < s->n_rings; r++) {
+ nfp_net_rx_ring_init(&rings[r], &nn->r_vecs[r], r);
if (nfp_net_rx_ring_alloc(&rings[r], fl_bufsz, s->dcnt))
goto err_free_prev;
@@ -1728,11 +1730,13 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
s->mtu = nn->netdev->mtu;
s->dcnt = nn->rxd_cnt;
s->rings = nn->rx_rings;
+ s->n_rings = nn->num_rx_rings;
nn->netdev->mtu = new.mtu;
nn->fl_bufsz = nfp_net_calc_fl_bufsz(nn, new.mtu);
nn->rxd_cnt = new.dcnt;
nn->rx_rings = new.rings;
+ nn->num_rx_rings = new.n_rings;
}
static void
@@ -1741,7 +1745,7 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
struct nfp_net_rx_ring *rings = s->rings;
unsigned int r;
- for (r = 0; r < nn->num_rx_rings; r++) {
+ for (r = 0; r < s->n_rings; r++) {
nfp_net_rx_ring_bufs_free(nn, &rings[r]);
nfp_net_rx_ring_free(&rings[r]);
}
@@ -1764,19 +1768,20 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
struct msix_entry *entry = &nn->irq_entries[r_vec->irq_idx];
int err;
+ /* Setup NAPI */
+ netif_napi_add(nn->netdev, &r_vec->napi,
+ nfp_net_poll, NAPI_POLL_WEIGHT);
+
snprintf(r_vec->name, sizeof(r_vec->name),
"%s-rxtx-%d", nn->netdev->name, idx);
err = request_irq(entry->vector, r_vec->handler, 0, r_vec->name, r_vec);
if (err) {
+ netif_napi_del(&r_vec->napi);
nn_err(nn, "Error requesting IRQ %d\n", entry->vector);
return err;
}
disable_irq(entry->vector);
- /* Setup NAPI */
- netif_napi_add(nn->netdev, &r_vec->napi,
- nfp_net_poll, NAPI_POLL_WEIGHT);
-
irq_set_affinity_hint(entry->vector, &r_vec->affinity_mask);
nn_dbg(nn, "RV%02d: irq=%03d/%03d\n", idx, entry->vector, entry->entry);
@@ -2036,10 +2041,12 @@ static int nfp_net_netdev_open(struct net_device *netdev)
{
struct nfp_net *nn = netdev_priv(netdev);
struct nfp_net_ring_set rx = {
+ .n_rings = nn->num_rx_rings,
.mtu = nn->netdev->mtu,
.dcnt = nn->rxd_cnt,
};
struct nfp_net_ring_set tx = {
+ .n_rings = nn->num_tx_rings,
.dcnt = nn->txd_cnt,
};
int err, r;
@@ -2239,49 +2246,89 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
}
static int
-nfp_net_ring_swap_enable(struct nfp_net *nn,
+nfp_net_ring_swap_enable(struct nfp_net *nn, unsigned int *num_vecs,
struct nfp_net_ring_set *rx,
struct nfp_net_ring_set *tx)
{
unsigned int r;
+ int err;
if (rx)
nfp_net_rx_ring_set_swap(nn, rx);
if (tx)
nfp_net_tx_ring_set_swap(nn, tx);
+ swap(*num_vecs, nn->num_r_vecs);
+
for (r = 0; r < nn->max_r_vecs; r++)
nfp_net_vector_assign_rings(nn, &nn->r_vecs[r], r);
+ if (nn->netdev->real_num_rx_queues != nn->num_rx_rings) {
+ if (!netif_is_rxfh_configured(nn->netdev))
+ nfp_net_rss_init_itbl(nn);
+
+ err = netif_set_real_num_rx_queues(nn->netdev,
+ nn->num_rx_rings);
+ if (err)
+ return err;
+ }
+
+ if (nn->netdev->real_num_tx_queues != nn->num_tx_rings) {
+ err = netif_set_real_num_tx_queues(nn->netdev,
+ nn->num_tx_rings);
+ if (err)
+ return err;
+ }
+
return __nfp_net_set_config_and_enable(nn);
}
static void
nfp_net_ring_reconfig_down(struct nfp_net *nn,
struct nfp_net_ring_set *rx,
- struct nfp_net_ring_set *tx)
+ struct nfp_net_ring_set *tx,
+ unsigned int num_vecs)
{
nn->netdev->mtu = rx ? rx->mtu : nn->netdev->mtu;
nn->fl_bufsz = nfp_net_calc_fl_bufsz(nn, nn->netdev->mtu);
nn->rxd_cnt = rx ? rx->dcnt : nn->rxd_cnt;
nn->txd_cnt = tx ? tx->dcnt : nn->txd_cnt;
+ nn->num_rx_rings = rx ? rx->n_rings : nn->num_rx_rings;
+ nn->num_tx_rings = tx ? tx->n_rings : nn->num_tx_rings;
+ nn->num_r_vecs = num_vecs;
+
+ if (!netif_is_rxfh_configured(nn->netdev))
+ nfp_net_rss_init_itbl(nn);
}
int
nfp_net_ring_reconfig(struct nfp_net *nn, struct nfp_net_ring_set *rx,
struct nfp_net_ring_set *tx)
{
+ unsigned int num_vecs, r;
int err;
+ num_vecs = max(rx ? rx->n_rings : nn->num_rx_rings,
+ tx ? tx->n_rings : nn->num_tx_rings);
+
if (!netif_running(nn->netdev)) {
- nfp_net_ring_reconfig_down(nn, rx, tx);
+ nfp_net_ring_reconfig_down(nn, rx, tx, num_vecs);
return 0;
}
/* Prepare new rings */
+ for (r = nn->num_r_vecs; r < num_vecs; r++) {
+ err = nfp_net_prepare_vector(nn, &nn->r_vecs[r], r);
+ if (err) {
+ num_vecs = r;
+ goto err_cleanup_vecs;
+ }
+ }
if (rx) {
- if (!nfp_net_rx_ring_set_prepare(nn, rx))
- return -ENOMEM;
+ if (!nfp_net_rx_ring_set_prepare(nn, rx)) {
+ err = -ENOMEM;
+ goto err_cleanup_vecs;
+ }
}
if (tx) {
if (!nfp_net_tx_ring_set_prepare(nn, tx)) {
@@ -2294,18 +2341,20 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
nfp_net_close_stack(nn);
nfp_net_clear_config_and_disable(nn);
- err = nfp_net_ring_swap_enable(nn, rx, tx);
+ err = nfp_net_ring_swap_enable(nn, &num_vecs, rx, tx);
if (err) {
int err2;
nfp_net_clear_config_and_disable(nn);
/* Try with old configuration and old rings */
- err2 = nfp_net_ring_swap_enable(nn, rx, tx);
+ err2 = nfp_net_ring_swap_enable(nn, &num_vecs, rx, tx);
if (err2)
nn_err(nn, "Can't restore ring config - FW communication failed (%d,%d)\n",
err, err2);
}
+ for (r = num_vecs - 1; r >= nn->num_r_vecs; r--)
+ nfp_net_cleanup_vector(nn, &nn->r_vecs[r]);
if (rx)
nfp_net_rx_ring_set_free(nn, rx);
@@ -2319,6 +2368,9 @@ static void nfp_net_rss_init_itbl(struct nfp_net *nn)
err_free_rx:
if (rx)
nfp_net_rx_ring_set_free(nn, rx);
+err_cleanup_vecs:
+ for (r = num_vecs - 1; r >= nn->num_r_vecs; r--)
+ nfp_net_cleanup_vector(nn, &nn->r_vecs[r]);
return err;
}
@@ -2326,6 +2378,7 @@ static int nfp_net_change_mtu(struct net_device *netdev, int new_mtu)
{
struct nfp_net *nn = netdev_priv(netdev);
struct nfp_net_ring_set rx = {
+ .n_rings = nn->num_rx_rings,
.mtu = new_mtu,
.dcnt = nn->rxd_cnt,
};
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
index 180cf70f0093..e90df8145323 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_debugfs.c
@@ -207,13 +207,13 @@ void nfp_net_debugfs_adapter_add(struct nfp_net *nn)
if (IS_ERR_OR_NULL(rx) || IS_ERR_OR_NULL(tx))
return;
- for (i = 0; i < nn->num_rx_rings; i++) {
+ for (i = 0; i < min(nn->max_rx_rings, nn->max_r_vecs); i++) {
sprintf(int_name, "%d", i);
debugfs_create_file(int_name, S_IRUSR, rx,
&nn->r_vecs[i], &nfp_rx_q_fops);
}
- for (i = 0; i < nn->num_tx_rings; i++) {
+ for (i = 0; i < min(nn->max_tx_rings, nn->max_r_vecs); i++) {
sprintf(int_name, "%d", i);
debugfs_create_file(int_name, S_IRUSR, tx,
&nn->r_vecs[i], &nfp_tx_q_fops);
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
index 3f48256dc03c..b87f1b73f200 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c
@@ -162,10 +162,12 @@ static int nfp_net_set_ring_size(struct nfp_net *nn, u32 rxd_cnt, u32 txd_cnt)
{
struct nfp_net_ring_set *reconfig_rx = NULL, *reconfig_tx = NULL;
struct nfp_net_ring_set rx = {
+ .n_rings = nn->num_rx_rings,
.mtu = nn->netdev->mtu,
.dcnt = rxd_cnt,
};
struct nfp_net_ring_set tx = {
+ .n_rings = nn->num_tx_rings,
.dcnt = txd_cnt,
};
@@ -648,6 +650,50 @@ static void nfp_net_get_channels(struct net_device *netdev,
channel->other_count = NFP_NET_NON_Q_VECTORS;
}
+static int nfp_net_set_num_rings(struct nfp_net *nn, unsigned int total_rx,
+ unsigned int total_tx)
+{
+ struct nfp_net_ring_set *reconfig_rx = NULL, *reconfig_tx = NULL;
+ struct nfp_net_ring_set rx = {
+ .n_rings = total_rx,
+ .mtu = nn->netdev->mtu,
+ .dcnt = nn->rxd_cnt,
+ };
+ struct nfp_net_ring_set tx = {
+ .n_rings = total_tx,
+ .dcnt = nn->txd_cnt,
+ };
+
+ if (nn->num_rx_rings != total_rx)
+ reconfig_rx = ℞
+ if (nn->num_tx_rings != total_tx)
+ reconfig_tx = &tx;
+
+ return nfp_net_ring_reconfig(nn, reconfig_rx, reconfig_tx);
+}
+
+static int nfp_net_set_channels(struct net_device *netdev,
+ struct ethtool_channels *channel)
+{
+ struct nfp_net *nn = netdev_priv(netdev);
+ unsigned int total_rx, total_tx;
+
+ /* Reject unsupported */
+ if (!channel->combined_count ||
+ channel->other_count != NFP_NET_NON_Q_VECTORS ||
+ (channel->rx_count && channel->tx_count))
+ return -EINVAL;
+
+ total_rx = channel->combined_count + channel->rx_count;
+ total_tx = channel->combined_count + channel->tx_count;
+
+ if (total_rx > min(nn->max_rx_rings, nn->max_r_vecs) ||
+ total_tx > min(nn->max_tx_rings, nn->max_r_vecs))
+ return -EINVAL;
+
+ return nfp_net_set_num_rings(nn, total_rx, total_tx);
+}
+
static const struct ethtool_ops nfp_net_ethtool_ops = {
.get_drvinfo = nfp_net_get_drvinfo,
.get_link = ethtool_op_get_link,
@@ -667,6 +713,7 @@ static void nfp_net_get_channels(struct net_device *netdev,
.get_coalesce = nfp_net_get_coalesce,
.set_coalesce = nfp_net_set_coalesce,
.get_channels = nfp_net_get_channels,
+ .set_channels = nfp_net_set_channels,
};
void nfp_net_set_ethtool_ops(struct net_device *netdev)
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 07/13] nfp: move RSS indirection table init into a separate function
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
We will need to rerun the initialization of the RSS indirection table
after the number of rings is changed. Move the code to a separate
function.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 15 ++++++++++-----
1 file changed, 10 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 2a4e1f1cb3c9..09cec6e2c6cf 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -2229,6 +2229,15 @@ static void nfp_net_set_rx_mode(struct net_device *netdev)
nn->ctrl = new_ctrl;
}
+static void nfp_net_rss_init_itbl(struct nfp_net *nn)
+{
+ int i;
+
+ for (i = 0; i < sizeof(nn->rss_itbl); i++)
+ nn->rss_itbl[i] =
+ ethtool_rxfh_indir_default(i, nn->num_rx_rings);
+}
+
static int
nfp_net_ring_swap_enable(struct nfp_net *nn,
struct nfp_net_ring_set *rx,
@@ -2707,13 +2716,9 @@ void nfp_net_netdev_free(struct nfp_net *nn)
*/
static void nfp_net_rss_init(struct nfp_net *nn)
{
- int i;
-
netdev_rss_key_fill(nn->rss_key, NFP_NET_CFG_RSS_KEY_SZ);
- for (i = 0; i < sizeof(nn->rss_itbl); i++)
- nn->rss_itbl[i] =
- ethtool_rxfh_indir_default(i, nn->num_rx_rings);
+ nfp_net_rss_init_itbl(nn);
/* Enable IPv4/IPv6 TCP by default */
nn->rss_cfg = NFP_NET_CFG_RSS_IPV4_TCP |
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 05/13] nfp: loosen relation between rings and IRQs vectors
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
Upcoming XDP support will break the assumption that one can iterate
over IRQ vectors to get to all the rings easily. Use nn->.x_ring
arrays directly.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
drivers/net/ethernet/netronome/nfp/nfp_net_common.c | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 50aeaea9e318..97cc21eae466 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -1920,9 +1920,9 @@ static void nfp_net_clear_config_and_disable(struct nfp_net *nn)
nn_err(nn, "Could not disable device: %d\n", err);
for (r = 0; r < nn->num_rx_rings; r++)
- nfp_net_rx_ring_reset(nn->r_vecs[r].rx_ring);
+ nfp_net_rx_ring_reset(&nn->rx_rings[r]);
for (r = 0; r < nn->num_tx_rings; r++)
- nfp_net_tx_ring_reset(nn, nn->r_vecs[r].tx_ring);
+ nfp_net_tx_ring_reset(nn, &nn->tx_rings[r]);
for (r = 0; r < nn->num_r_vecs; r++)
nfp_net_vec_clear_ring_data(nn, r);
@@ -2000,7 +2000,7 @@ static int __nfp_net_set_config_and_enable(struct nfp_net *nn)
nn->ctrl = new_ctrl;
for (r = 0; r < nn->num_rx_rings; r++)
- nfp_net_rx_ring_fill_freelist(nn->r_vecs[r].rx_ring);
+ nfp_net_rx_ring_fill_freelist(&nn->rx_rings[r]);
/* Since reconfiguration requests while NFP is down are ignored we
* have to wipe the entire VXLAN configuration and reinitialize it.
@@ -2173,11 +2173,11 @@ static void nfp_net_close_free_all(struct nfp_net *nn)
unsigned int r;
for (r = 0; r < nn->num_rx_rings; r++) {
- nfp_net_rx_ring_bufs_free(nn, nn->r_vecs[r].rx_ring);
- nfp_net_rx_ring_free(nn->r_vecs[r].rx_ring);
+ nfp_net_rx_ring_bufs_free(nn, &nn->rx_rings[r]);
+ nfp_net_rx_ring_free(&nn->rx_rings[r]);
}
for (r = 0; r < nn->num_tx_rings; r++)
- nfp_net_tx_ring_free(nn->r_vecs[r].tx_ring);
+ nfp_net_tx_ring_free(&nn->tx_rings[r]);
for (r = 0; r < nn->num_r_vecs; r++)
nfp_net_cleanup_vector(nn, &nn->r_vecs[r]);
--
1.9.1
^ permalink raw reply related
* [PATCH net-next resend 06/13] nfp: add helper to reassign rings to IRQ vectors
From: Jakub Kicinski @ 2016-11-03 17:12 UTC (permalink / raw)
To: netdev; +Cc: Jakub Kicinski
In-Reply-To: <1478193129-23476-1-git-send-email-jakub.kicinski@netronome.com>
Instead of fixing ring -> vector relations up in ring swap functions
put the reassignment into a helper function which will reinit all
links.
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
---
.../net/ethernet/netronome/nfp/nfp_net_common.c | 40 +++++++++-------------
1 file changed, 16 insertions(+), 24 deletions(-)
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
index 97cc21eae466..2a4e1f1cb3c9 100644
--- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
+++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c
@@ -1601,16 +1601,11 @@ static int nfp_net_tx_ring_alloc(struct nfp_net_tx_ring *tx_ring, u32 cnt)
static void
nfp_net_tx_ring_set_swap(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
- struct nfp_net_tx_ring *rings = s->rings;
struct nfp_net_ring_set new = *s;
- unsigned int r;
s->dcnt = nn->txd_cnt;
s->rings = nn->tx_rings;
- for (r = 0; r < nn->num_tx_rings; r++)
- nn->tx_rings[r].r_vec->tx_ring = &rings[r];
-
nn->txd_cnt = new.dcnt;
nn->tx_rings = new.rings;
}
@@ -1728,17 +1723,12 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
static void
nfp_net_rx_ring_set_swap(struct nfp_net *nn, struct nfp_net_ring_set *s)
{
- struct nfp_net_rx_ring *rings = s->rings;
struct nfp_net_ring_set new = *s;
- unsigned int r;
s->mtu = nn->netdev->mtu;
s->dcnt = nn->rxd_cnt;
s->rings = nn->rx_rings;
- for (r = 0; r < nn->num_rx_rings; r++)
- nn->rx_rings[r].r_vec->rx_ring = &rings[r];
-
nn->netdev->mtu = new.mtu;
nn->fl_bufsz = nfp_net_calc_fl_bufsz(nn, new.mtu);
nn->rxd_cnt = new.dcnt;
@@ -1759,6 +1749,14 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
kfree(rings);
}
+static void
+nfp_net_vector_assign_rings(struct nfp_net *nn, struct nfp_net_r_vector *r_vec,
+ int idx)
+{
+ r_vec->rx_ring = idx < nn->num_rx_rings ? &nn->rx_rings[idx] : NULL;
+ r_vec->tx_ring = idx < nn->num_tx_rings ? &nn->tx_rings[idx] : NULL;
+}
+
static int
nfp_net_prepare_vector(struct nfp_net *nn, struct nfp_net_r_vector *r_vec,
int idx)
@@ -1766,20 +1764,6 @@ static void nfp_net_rx_ring_free(struct nfp_net_rx_ring *rx_ring)
struct msix_entry *entry = &nn->irq_entries[r_vec->irq_idx];
int err;
- if (idx < nn->num_tx_rings) {
- r_vec->tx_ring = &nn->tx_rings[idx];
- nfp_net_tx_ring_init(r_vec->tx_ring, r_vec, idx);
- } else {
- r_vec->tx_ring = NULL;
- }
-
- if (idx < nn->num_rx_rings) {
- r_vec->rx_ring = &nn->rx_rings[idx];
- nfp_net_rx_ring_init(r_vec->rx_ring, r_vec, idx);
- } else {
- r_vec->rx_ring = NULL;
- }
-
snprintf(r_vec->name, sizeof(r_vec->name),
"%s-rxtx-%d", nn->netdev->name, idx);
err = request_irq(entry->vector, r_vec->handler, 0, r_vec->name, r_vec);
@@ -2100,6 +2084,9 @@ static int nfp_net_netdev_open(struct net_device *netdev)
goto err_free_rx_rings;
}
+ for (r = 0; r < nn->max_r_vecs; r++)
+ nfp_net_vector_assign_rings(nn, &nn->r_vecs[r], r);
+
err = netif_set_real_num_tx_queues(netdev, nn->num_tx_rings);
if (err)
goto err_free_rings;
@@ -2247,11 +2234,16 @@ static void nfp_net_set_rx_mode(struct net_device *netdev)
struct nfp_net_ring_set *rx,
struct nfp_net_ring_set *tx)
{
+ unsigned int r;
+
if (rx)
nfp_net_rx_ring_set_swap(nn, rx);
if (tx)
nfp_net_tx_ring_set_swap(nn, tx);
+ for (r = 0; r < nn->max_r_vecs; r++)
+ nfp_net_vector_assign_rings(nn, &nn->r_vecs[r], r);
+
return __nfp_net_set_config_and_enable(nn);
}
--
1.9.1
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox