summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--Documentation/devicetree/bindings/net/ti,icssg-prueth.yaml1
-rw-r--r--Documentation/networking/ip-sysctl.rst7
-rw-r--r--drivers/net/ethernet/ti/icssg/icssg_prueth.c5
-rw-r--r--drivers/net/wwan/wwan_core.c5
-rw-r--r--include/linux/tcp.h14
-rw-r--r--include/net/dst.h4
-rw-r--r--include/net/netns/ipv4.h1
-rw-r--r--include/net/sock.h9
-rw-r--r--net/core/dst.c10
-rw-r--r--net/core/sock.c6
-rw-r--r--net/ipv4/route.c6
-rw-r--r--net/ipv4/sysctl_net_ipv4.c9
-rw-r--r--net/ipv4/tcp_input.c8
-rw-r--r--net/ipv4/tcp_ipv4.c1
-rw-r--r--net/ipv4/tcp_output.c15
-rw-r--r--net/ipv6/route.c4
-rw-r--r--net/openvswitch/actions.c4
-rw-r--r--net/sched/sch_frag.c4
-rw-r--r--net/xfrm/xfrm_policy.c2
19 files changed, 71 insertions, 44 deletions
diff --git a/Documentation/devicetree/bindings/net/ti,icssg-prueth.yaml b/Documentation/devicetree/bindings/net/ti,icssg-prueth.yaml
index 311c570165f9..836d2d60e87d 100644
--- a/Documentation/devicetree/bindings/net/ti,icssg-prueth.yaml
+++ b/Documentation/devicetree/bindings/net/ti,icssg-prueth.yaml
@@ -19,6 +19,7 @@ allOf:
properties:
compatible:
enum:
+ - ti,am642-icssg-prueth # for AM64x SoC family
- ti,am654-icssg-prueth # for AM65x SoC family
sram:
diff --git a/Documentation/networking/ip-sysctl.rst b/Documentation/networking/ip-sysctl.rst
index a66054d0763a..5bfa1837968c 100644
--- a/Documentation/networking/ip-sysctl.rst
+++ b/Documentation/networking/ip-sysctl.rst
@@ -745,6 +745,13 @@ tcp_comp_sack_nr - INTEGER
Default : 44
+tcp_backlog_ack_defer - BOOLEAN
+ If set, user thread processing socket backlog tries sending
+ one ACK for the whole queue. This helps to avoid potential
+ long latencies at end of a TCP socket syscall.
+
+ Default : true
+
tcp_slow_start_after_idle - BOOLEAN
If set, provide RFC2861 behavior and time out the congestion
window after an idle period. An idle period is defined at
diff --git a/drivers/net/ethernet/ti/icssg/icssg_prueth.c b/drivers/net/ethernet/ti/icssg/icssg_prueth.c
index 410612f43cbd..92b13057d4de 100644
--- a/drivers/net/ethernet/ti/icssg/icssg_prueth.c
+++ b/drivers/net/ethernet/ti/icssg/icssg_prueth.c
@@ -2313,8 +2313,13 @@ static const struct prueth_pdata am654_icssg_pdata = {
.quirk_10m_link_issue = 1,
};
+static const struct prueth_pdata am64x_icssg_pdata = {
+ .fdqring_mode = K3_RINGACC_RING_MODE_RING,
+};
+
static const struct of_device_id prueth_dt_match[] = {
{ .compatible = "ti,am654-icssg-prueth", .data = &am654_icssg_pdata },
+ { .compatible = "ti,am642-icssg-prueth", .data = &am64x_icssg_pdata },
{ /* sentinel */ }
};
MODULE_DEVICE_TABLE(of, prueth_dt_match);
diff --git a/drivers/net/wwan/wwan_core.c b/drivers/net/wwan/wwan_core.c
index 284ab1f56391..87df60916960 100644
--- a/drivers/net/wwan/wwan_core.c
+++ b/drivers/net/wwan/wwan_core.c
@@ -1,6 +1,7 @@
// SPDX-License-Identifier: GPL-2.0-only
/* Copyright (c) 2021, Linaro Ltd <loic.poulain@linaro.org> */
+#include <linux/bitmap.h>
#include <linux/err.h>
#include <linux/errno.h>
#include <linux/debugfs.h>
@@ -395,7 +396,7 @@ static int __wwan_port_dev_assign_name(struct wwan_port *port, const char *fmt)
char buf[0x20];
int id;
- idmap = (unsigned long *)get_zeroed_page(GFP_KERNEL);
+ idmap = bitmap_zalloc(max_ports, GFP_KERNEL);
if (!idmap)
return -ENOMEM;
@@ -414,7 +415,7 @@ static int __wwan_port_dev_assign_name(struct wwan_port *port, const char *fmt)
/* Allocate unique id */
id = find_first_zero_bit(idmap, max_ports);
- free_page((unsigned long)idmap);
+ bitmap_free(idmap);
snprintf(buf, sizeof(buf), fmt, id); /* Name generation */
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 3c5efeeb024f..44d946161d4a 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -463,15 +463,17 @@ enum tsq_enum {
TCP_MTU_REDUCED_DEFERRED, /* tcp_v{4|6}_err() could not call
* tcp_v{4|6}_mtu_reduced()
*/
+ TCP_ACK_DEFERRED, /* TX pure ack is deferred */
};
enum tsq_flags {
- TSQF_THROTTLED = (1UL << TSQ_THROTTLED),
- TSQF_QUEUED = (1UL << TSQ_QUEUED),
- TCPF_TSQ_DEFERRED = (1UL << TCP_TSQ_DEFERRED),
- TCPF_WRITE_TIMER_DEFERRED = (1UL << TCP_WRITE_TIMER_DEFERRED),
- TCPF_DELACK_TIMER_DEFERRED = (1UL << TCP_DELACK_TIMER_DEFERRED),
- TCPF_MTU_REDUCED_DEFERRED = (1UL << TCP_MTU_REDUCED_DEFERRED),
+ TSQF_THROTTLED = BIT(TSQ_THROTTLED),
+ TSQF_QUEUED = BIT(TSQ_QUEUED),
+ TCPF_TSQ_DEFERRED = BIT(TCP_TSQ_DEFERRED),
+ TCPF_WRITE_TIMER_DEFERRED = BIT(TCP_WRITE_TIMER_DEFERRED),
+ TCPF_DELACK_TIMER_DEFERRED = BIT(TCP_DELACK_TIMER_DEFERRED),
+ TCPF_MTU_REDUCED_DEFERRED = BIT(TCP_MTU_REDUCED_DEFERRED),
+ TCPF_ACK_DEFERRED = BIT(TCP_ACK_DEFERRED),
};
#define tcp_sk(ptr) container_of_const(ptr, struct tcp_sock, inet_conn.icsk_inet.sk)
diff --git a/include/net/dst.h b/include/net/dst.h
index 78884429deed..f8b8599a0600 100644
--- a/include/net/dst.h
+++ b/include/net/dst.h
@@ -392,10 +392,10 @@ static inline int dst_discard(struct sk_buff *skb)
{
return dst_discard_out(&init_net, skb->sk, skb);
}
-void *dst_alloc(struct dst_ops *ops, struct net_device *dev, int initial_ref,
+void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
int initial_obsolete, unsigned short flags);
void dst_init(struct dst_entry *dst, struct dst_ops *ops,
- struct net_device *dev, int initial_ref, int initial_obsolete,
+ struct net_device *dev, int initial_obsolete,
unsigned short flags);
struct dst_entry *dst_destroy(struct dst_entry *dst);
void dst_dev_put(struct dst_entry *dst);
diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index 7a41c4791536..d96d05b08819 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -132,6 +132,7 @@ struct netns_ipv4 {
u8 sysctl_tcp_syncookies;
u8 sysctl_tcp_migrate_req;
u8 sysctl_tcp_comp_sack_nr;
+ u8 sysctl_tcp_backlog_ack_defer;
int sysctl_tcp_reordering;
u8 sysctl_tcp_retries1;
u8 sysctl_tcp_retries2;
diff --git a/include/net/sock.h b/include/net/sock.h
index b770261fbdaf..676146e9d181 100644
--- a/include/net/sock.h
+++ b/include/net/sock.h
@@ -1823,12 +1823,11 @@ static inline bool sock_owned_by_user_nocheck(const struct sock *sk)
static inline void sock_release_ownership(struct sock *sk)
{
- if (sock_owned_by_user_nocheck(sk)) {
- sk->sk_lock.owned = 0;
+ DEBUG_NET_WARN_ON_ONCE(!sock_owned_by_user_nocheck(sk));
+ sk->sk_lock.owned = 0;
- /* The sk_lock has mutex_unlock() semantics: */
- mutex_release(&sk->sk_lock.dep_map, _RET_IP_);
- }
+ /* The sk_lock has mutex_unlock() semantics: */
+ mutex_release(&sk->sk_lock.dep_map, _RET_IP_);
}
/* no reclassification while locks are held */
diff --git a/net/core/dst.c b/net/core/dst.c
index 980e2fd2f013..6838d3212c37 100644
--- a/net/core/dst.c
+++ b/net/core/dst.c
@@ -45,7 +45,7 @@ const struct dst_metrics dst_default_metrics = {
EXPORT_SYMBOL(dst_default_metrics);
void dst_init(struct dst_entry *dst, struct dst_ops *ops,
- struct net_device *dev, int initial_ref, int initial_obsolete,
+ struct net_device *dev, int initial_obsolete,
unsigned short flags)
{
dst->dev = dev;
@@ -66,7 +66,7 @@ void dst_init(struct dst_entry *dst, struct dst_ops *ops,
dst->tclassid = 0;
#endif
dst->lwtstate = NULL;
- rcuref_init(&dst->__rcuref, initial_ref);
+ rcuref_init(&dst->__rcuref, 1);
INIT_LIST_HEAD(&dst->rt_uncached);
dst->__use = 0;
dst->lastuse = jiffies;
@@ -77,7 +77,7 @@ void dst_init(struct dst_entry *dst, struct dst_ops *ops,
EXPORT_SYMBOL(dst_init);
void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
- int initial_ref, int initial_obsolete, unsigned short flags)
+ int initial_obsolete, unsigned short flags)
{
struct dst_entry *dst;
@@ -90,7 +90,7 @@ void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
if (!dst)
return NULL;
- dst_init(dst, ops, dev, initial_ref, initial_obsolete, flags);
+ dst_init(dst, ops, dev, initial_obsolete, flags);
return dst;
}
@@ -270,7 +270,7 @@ static void __metadata_dst_init(struct metadata_dst *md_dst,
struct dst_entry *dst;
dst = &md_dst->dst;
- dst_init(dst, &dst_blackhole_ops, NULL, 1, DST_OBSOLETE_NONE,
+ dst_init(dst, &dst_blackhole_ops, NULL, DST_OBSOLETE_NONE,
DST_METADATA | DST_NOCOUNT);
memset(dst + 1, 0, sizeof(*md_dst) + optslen - sizeof(*dst));
md_dst->type = type;
diff --git a/net/core/sock.c b/net/core/sock.c
index 16584e2dd648..bb89b88bc1e8 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -3001,6 +3001,9 @@ void __sk_flush_backlog(struct sock *sk)
{
spin_lock_bh(&sk->sk_lock.slock);
__release_sock(sk);
+
+ if (sk->sk_prot->release_cb)
+ sk->sk_prot->release_cb(sk);
spin_unlock_bh(&sk->sk_lock.slock);
}
EXPORT_SYMBOL_GPL(__sk_flush_backlog);
@@ -3519,9 +3522,6 @@ void release_sock(struct sock *sk)
if (sk->sk_backlog.tail)
__release_sock(sk);
- /* Warning : release_cb() might need to release sk ownership,
- * ie call sock_release_ownership(sk) before us.
- */
if (sk->sk_prot->release_cb)
sk->sk_prot->release_cb(sk);
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 66f419e7f9a7..fb3045692b99 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1630,7 +1630,7 @@ struct rtable *rt_dst_alloc(struct net_device *dev,
{
struct rtable *rt;
- rt = dst_alloc(&ipv4_dst_ops, dev, 1, DST_OBSOLETE_FORCE_CHK,
+ rt = dst_alloc(&ipv4_dst_ops, dev, DST_OBSOLETE_FORCE_CHK,
(noxfrm ? DST_NOXFRM : 0));
if (rt) {
@@ -1658,7 +1658,7 @@ struct rtable *rt_dst_clone(struct net_device *dev, struct rtable *rt)
{
struct rtable *new_rt;
- new_rt = dst_alloc(&ipv4_dst_ops, dev, 1, DST_OBSOLETE_FORCE_CHK,
+ new_rt = dst_alloc(&ipv4_dst_ops, dev, DST_OBSOLETE_FORCE_CHK,
rt->dst.flags);
if (new_rt) {
@@ -2832,7 +2832,7 @@ struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_or
struct rtable *ort = (struct rtable *) dst_orig;
struct rtable *rt;
- rt = dst_alloc(&ipv4_dst_blackhole_ops, NULL, 1, DST_OBSOLETE_DEAD, 0);
+ rt = dst_alloc(&ipv4_dst_blackhole_ops, NULL, DST_OBSOLETE_DEAD, 0);
if (rt) {
struct dst_entry *new = &rt->dst;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 6ac890b4073f..e7f024d93572 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -1367,6 +1367,15 @@ static struct ctl_table ipv4_net_table[] = {
.extra1 = SYSCTL_ZERO,
},
{
+ .procname = "tcp_backlog_ack_defer",
+ .data = &init_net.ipv4.sysctl_tcp_backlog_ack_defer,
+ .maxlen = sizeof(u8),
+ .mode = 0644,
+ .proc_handler = proc_dou8vec_minmax,
+ .extra1 = SYSCTL_ZERO,
+ .extra2 = SYSCTL_ONE,
+ },
+ {
.procname = "tcp_reflect_tos",
.data = &init_net.ipv4.sysctl_tcp_reflect_tos,
.maxlen = sizeof(u8),
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 06fe1cf645d5..41b471748437 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -5553,6 +5553,14 @@ static void __tcp_ack_snd_check(struct sock *sk, int ofo_possible)
tcp_in_quickack_mode(sk) ||
/* Protocol state mandates a one-time immediate ACK */
inet_csk(sk)->icsk_ack.pending & ICSK_ACK_NOW) {
+ /* If we are running from __release_sock() in user context,
+ * Defer the ack until tcp_release_cb().
+ */
+ if (sock_owned_by_user_nocheck(sk) &&
+ READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_backlog_ack_defer)) {
+ set_bit(TCP_ACK_DEFERRED, &sk->sk_tsq_flags);
+ return;
+ }
send_now:
tcp_send_ack(sk);
return;
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 27140e5cdc06..f13eb7e23d03 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -3263,6 +3263,7 @@ static int __net_init tcp_sk_init(struct net *net)
net->ipv4.sysctl_tcp_comp_sack_delay_ns = NSEC_PER_MSEC;
net->ipv4.sysctl_tcp_comp_sack_slack_ns = 100 * NSEC_PER_USEC;
net->ipv4.sysctl_tcp_comp_sack_nr = 44;
+ net->ipv4.sysctl_tcp_backlog_ack_defer = 1;
net->ipv4.sysctl_tcp_fastopen = TFO_CLIENT_ENABLE;
net->ipv4.sysctl_tcp_fastopen_blackhole_timeout = 0;
atomic_set(&net->ipv4.tfo_active_disable_times, 0);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index ccfc8bbf7455..1fc1f879cfd6 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1077,7 +1077,8 @@ static void tcp_tasklet_func(struct tasklet_struct *t)
#define TCP_DEFERRED_ALL (TCPF_TSQ_DEFERRED | \
TCPF_WRITE_TIMER_DEFERRED | \
TCPF_DELACK_TIMER_DEFERRED | \
- TCPF_MTU_REDUCED_DEFERRED)
+ TCPF_MTU_REDUCED_DEFERRED | \
+ TCPF_ACK_DEFERRED)
/**
* tcp_release_cb - tcp release_sock() callback
* @sk: socket
@@ -1101,16 +1102,6 @@ void tcp_release_cb(struct sock *sk)
tcp_tsq_write(sk);
__sock_put(sk);
}
- /* Here begins the tricky part :
- * We are called from release_sock() with :
- * 1) BH disabled
- * 2) sk_lock.slock spinlock held
- * 3) socket owned by us (sk->sk_lock.owned == 1)
- *
- * But following code is meant to be called from BH handlers,
- * so we should keep BH disabled, but early release socket ownership
- */
- sock_release_ownership(sk);
if (flags & TCPF_WRITE_TIMER_DEFERRED) {
tcp_write_timer_handler(sk);
@@ -1124,6 +1115,8 @@ void tcp_release_cb(struct sock *sk)
inet_csk(sk)->icsk_af_ops->mtu_reduced(sk);
__sock_put(sk);
}
+ if ((flags & TCPF_ACK_DEFERRED) && inet_csk_ack_scheduled(sk))
+ tcp_send_ack(sk);
}
EXPORT_SYMBOL(tcp_release_cb);
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 9c687b357e6a..9d8dfc7423e4 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -341,7 +341,7 @@ struct rt6_info *ip6_dst_alloc(struct net *net, struct net_device *dev,
int flags)
{
struct rt6_info *rt = dst_alloc(&net->ipv6.ip6_dst_ops, dev,
- 1, DST_OBSOLETE_FORCE_CHK, flags);
+ DST_OBSOLETE_FORCE_CHK, flags);
if (rt) {
rt6_info_init(rt);
@@ -2655,7 +2655,7 @@ struct dst_entry *ip6_blackhole_route(struct net *net, struct dst_entry *dst_ori
struct net_device *loopback_dev = net->loopback_dev;
struct dst_entry *new = NULL;
- rt = dst_alloc(&ip6_dst_blackhole_ops, loopback_dev, 1,
+ rt = dst_alloc(&ip6_dst_blackhole_ops, loopback_dev,
DST_OBSOLETE_DEAD, 0);
if (rt) {
rt6_info_init(rt);
diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c
index fd66014d8a76..5f8094acd056 100644
--- a/net/openvswitch/actions.c
+++ b/net/openvswitch/actions.c
@@ -873,7 +873,7 @@ static void ovs_fragment(struct net *net, struct vport *vport,
prepare_frag(vport, skb, orig_network_offset,
ovs_key_mac_proto(key));
- dst_init(&ovs_rt.dst, &ovs_dst_ops, NULL, 1,
+ dst_init(&ovs_rt.dst, &ovs_dst_ops, NULL,
DST_OBSOLETE_NONE, DST_NOCOUNT);
ovs_rt.dst.dev = vport->dev;
@@ -890,7 +890,7 @@ static void ovs_fragment(struct net *net, struct vport *vport,
prepare_frag(vport, skb, orig_network_offset,
ovs_key_mac_proto(key));
memset(&ovs_rt, 0, sizeof(ovs_rt));
- dst_init(&ovs_rt.dst, &ovs_dst_ops, NULL, 1,
+ dst_init(&ovs_rt.dst, &ovs_dst_ops, NULL,
DST_OBSOLETE_NONE, DST_NOCOUNT);
ovs_rt.dst.dev = vport->dev;
diff --git a/net/sched/sch_frag.c b/net/sched/sch_frag.c
index a9bd0a235890..ce63414185fd 100644
--- a/net/sched/sch_frag.c
+++ b/net/sched/sch_frag.c
@@ -96,7 +96,7 @@ static int sch_fragment(struct net *net, struct sk_buff *skb,
unsigned long orig_dst;
sch_frag_prepare_frag(skb, xmit);
- dst_init(&sch_frag_rt.dst, &sch_frag_dst_ops, NULL, 1,
+ dst_init(&sch_frag_rt.dst, &sch_frag_dst_ops, NULL,
DST_OBSOLETE_NONE, DST_NOCOUNT);
sch_frag_rt.dst.dev = skb->dev;
@@ -112,7 +112,7 @@ static int sch_fragment(struct net *net, struct sk_buff *skb,
sch_frag_prepare_frag(skb, xmit);
memset(&sch_frag_rt, 0, sizeof(sch_frag_rt));
- dst_init(&sch_frag_rt.dst, &sch_frag_dst_ops, NULL, 1,
+ dst_init(&sch_frag_rt.dst, &sch_frag_dst_ops, NULL,
DST_OBSOLETE_NONE, DST_NOCOUNT);
sch_frag_rt.dst.dev = skb->dev;
diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index d6b405782b63..c4c4fc29ccf5 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -2561,7 +2561,7 @@ static inline struct xfrm_dst *xfrm_alloc_dst(struct net *net, int family)
default:
BUG();
}
- xdst = dst_alloc(dst_ops, NULL, 1, DST_OBSOLETE_NONE, 0);
+ xdst = dst_alloc(dst_ops, NULL, DST_OBSOLETE_NONE, 0);
if (likely(xdst)) {
memset_after(xdst, 0, u.dst);