From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: [PATCH net-next 0/6] tcp: better handling of memory pressure Date: Fri, 15 May 2015 07:53:52 -0700 Message-ID: <1431701638-24451-1-git-send-email-edumazet@google.com> Cc: netdev , Jason Baron , Neal Cardwell , Yuchung Cheng , Eric Dumazet , Eric Dumazet To: "David S. Miller" Return-path: Received: from mail-yk0-f201.google.com ([209.85.160.201]:33743 "EHLO mail-yk0-f201.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754101AbbEOPCG (ORCPT ); Fri, 15 May 2015 11:02:06 -0400 Received: by ykp9 with SMTP id 9so3220957ykp.0 for ; Fri, 15 May 2015 08:02:05 -0700 (PDT) Sender: netdev-owner@vger.kernel.org List-ID: When testing commit 790ba4566c1a ("tcp: set SOCK_NOSPACE under memory pressure") using edge triggered epoll applications, I found various issues under memory pressure and thousands of active sockets. This patch series is a first round to solve these issues, in send and receive paths. There are probably other fixes needed, but with this series, my tests now all succeed. Eric Dumazet (6): net: fix sk_mem_reclaim_partial() tcp: rename sk_forced_wmem_schedule() to sk_forced_mem_schedule() tcp: introduce tcp_under_memory_pressure() tcp: fix behavior for epoll edge trigger tcp: allow one skb to be received per socket under memory pressure tcp: halves tcp_mem[] limits include/net/sock.h | 6 +++--- include/net/tcp.h | 10 ++++++++++ net/core/sock.c | 9 +++++---- net/ipv4/tcp.c | 24 ++++++++++++++++++------ net/ipv4/tcp_input.c | 18 ++++++++++-------- net/ipv4/tcp_output.c | 10 ++++++---- net/ipv4/tcp_timer.c | 2 +- 7 files changed, 53 insertions(+), 26 deletions(-) -- 2.2.0.rc0.207.ga3a616c