From: Eric Dumazet <edumazet@google.com>
To: "David S. Miller" <davem@davemloft.net>
Cc: netdev <netdev@vger.kernel.org>, Jason Baron <jbaron@akamai.com>,
Neal Cardwell <ncardwell@google.com>,
Yuchung Cheng <ycheng@google.com>,
Eric Dumazet <eric.dumazet@gmail.com>,
Eric Dumazet <edumazet@google.com>
Subject: [PATCH net-next 3/6] tcp: introduce tcp_under_memory_pressure()
Date: Fri, 15 May 2015 07:53:55 -0700 [thread overview]
Message-ID: <1431701638-24451-4-git-send-email-edumazet@google.com> (raw)
In-Reply-To: <1431701638-24451-1-git-send-email-edumazet@google.com>
Introduce an optimized version of sk_under_memory_pressure()
for TCP. Our intent is to use it in fast paths.
Signed-off-by: Eric Dumazet <edumazet@google.com>
---
include/net/tcp.h | 8 ++++++++
net/ipv4/tcp_input.c | 8 ++++----
net/ipv4/tcp_output.c | 4 ++--
net/ipv4/tcp_timer.c | 2 +-
4 files changed, 15 insertions(+), 7 deletions(-)
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 841691a296dc..0d85223efa4c 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -286,6 +286,14 @@ extern atomic_long_t tcp_memory_allocated;
extern struct percpu_counter tcp_sockets_allocated;
extern int tcp_memory_pressure;
+/* optimized version of sk_under_memory_pressure() for TCP sockets */
+static inline bool tcp_under_memory_pressure(const struct sock *sk)
+{
+ if (mem_cgroup_sockets_enabled && sk->sk_cgrp)
+ return !!sk->sk_cgrp->memory_pressure;
+
+ return tcp_memory_pressure;
+}
/*
* The next routines deal with comparing 32 bit unsigned ints
* and worry about wraparound (automatic with unsigned arithmetic).
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index cf8b20ff6658..093779f7e893 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -359,7 +359,7 @@ static void tcp_grow_window(struct sock *sk, const struct sk_buff *skb)
/* Check #1 */
if (tp->rcv_ssthresh < tp->window_clamp &&
(int)tp->rcv_ssthresh < tcp_space(sk) &&
- !sk_under_memory_pressure(sk)) {
+ !tcp_under_memory_pressure(sk)) {
int incr;
/* Check #2. Increase window, if skb with such overhead
@@ -446,7 +446,7 @@ static void tcp_clamp_window(struct sock *sk)
if (sk->sk_rcvbuf < sysctl_tcp_rmem[2] &&
!(sk->sk_userlocks & SOCK_RCVBUF_LOCK) &&
- !sk_under_memory_pressure(sk) &&
+ !tcp_under_memory_pressure(sk) &&
sk_memory_allocated(sk) < sk_prot_mem_limits(sk, 0)) {
sk->sk_rcvbuf = min(atomic_read(&sk->sk_rmem_alloc),
sysctl_tcp_rmem[2]);
@@ -4781,7 +4781,7 @@ static int tcp_prune_queue(struct sock *sk)
if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
tcp_clamp_window(sk);
- else if (sk_under_memory_pressure(sk))
+ else if (tcp_under_memory_pressure(sk))
tp->rcv_ssthresh = min(tp->rcv_ssthresh, 4U * tp->advmss);
tcp_collapse_ofo_queue(sk);
@@ -4825,7 +4825,7 @@ static bool tcp_should_expand_sndbuf(const struct sock *sk)
return false;
/* If we are under global TCP memory pressure, do not expand. */
- if (sk_under_memory_pressure(sk))
+ if (tcp_under_memory_pressure(sk))
return false;
/* If we are under soft global TCP memory pressure, do not expand. */
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index bac1a950d087..08c2cc40b26d 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2392,7 +2392,7 @@ u32 __tcp_select_window(struct sock *sk)
if (free_space < (full_space >> 1)) {
icsk->icsk_ack.quick = 0;
- if (sk_under_memory_pressure(sk))
+ if (tcp_under_memory_pressure(sk))
tp->rcv_ssthresh = min(tp->rcv_ssthresh,
4U * tp->advmss);
@@ -2843,7 +2843,7 @@ void tcp_send_fin(struct sock *sk)
* Note: in the latter case, FIN packet will be sent after a timeout,
* as TCP stack thinks it has already been transmitted.
*/
- if (tskb && (tcp_send_head(sk) || sk_under_memory_pressure(sk))) {
+ if (tskb && (tcp_send_head(sk) || tcp_under_memory_pressure(sk))) {
coalesce:
TCP_SKB_CB(tskb)->tcp_flags |= TCPHDR_FIN;
TCP_SKB_CB(tskb)->end_seq++;
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 65bf670e8714..5b752f58a900 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -247,7 +247,7 @@ void tcp_delack_timer_handler(struct sock *sk)
}
out:
- if (sk_under_memory_pressure(sk))
+ if (tcp_under_memory_pressure(sk))
sk_mem_reclaim(sk);
}
--
2.2.0.rc0.207.ga3a616c
next prev parent reply other threads:[~2015-05-15 14:54 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-15 14:53 [PATCH net-next 0/6] tcp: better handling of memory pressure Eric Dumazet
2015-05-15 14:53 ` [PATCH net-next 1/6] net: fix sk_mem_reclaim_partial() Eric Dumazet
2015-05-15 14:53 ` [PATCH net-next 2/6] tcp: rename sk_forced_wmem_schedule() to sk_forced_mem_schedule() Eric Dumazet
2015-05-15 14:53 ` Eric Dumazet [this message]
2015-05-15 14:53 ` [PATCH net-next 4/6] tcp: fix behavior for epoll edge trigger Eric Dumazet
2015-05-15 14:53 ` [PATCH net-next 5/6] tcp: allow one skb to be received per socket under memory pressure Eric Dumazet
2015-05-15 18:20 ` Jason Baron
2015-05-15 18:22 ` Eric Dumazet
2015-05-15 14:53 ` [PATCH net-next 6/6] tcp: halves tcp_mem[] limits Eric Dumazet
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1431701638-24451-4-git-send-email-edumazet@google.com \
--to=edumazet@google.com \
--cc=davem@davemloft.net \
--cc=eric.dumazet@gmail.com \
--cc=jbaron@akamai.com \
--cc=ncardwell@google.com \
--cc=netdev@vger.kernel.org \
--cc=ycheng@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox