From d8b81175e412c7abebdb5b37d8a84d5fd19b1aad Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 22 Sep 2021 19:26:43 +0200 Subject: tcp: remove sk_{tr}x_skb_cache This reverts the following patches : - commit 2e05fcae83c4 ("tcp: fix compile error if !CONFIG_SYSCTL") - commit 4f661542a402 ("tcp: fix zerocopy and notsent_lowat issues") - commit 472c2e07eef0 ("tcp: add one skb cache for tx") - commit 8b27dae5a2e8 ("tcp: add one skb cache for rx") Having a cache of one skb (in each direction) per TCP socket is fragile, since it can cause a significant increase of memory needs, and not good enough for high speed flows anyway where more than one skb is needed. We want instead to add a generic infrastructure, with more flexible per-cpu caches, for alien NUMA nodes. Acked-by: Paolo Abeni Acked-by: Mat Martineau Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- Documentation/networking/ip-sysctl.rst | 8 -------- 1 file changed, 8 deletions(-) (limited to 'Documentation/networking/ip-sysctl.rst') diff --git a/Documentation/networking/ip-sysctl.rst b/Documentation/networking/ip-sysctl.rst index d91ab28718d4..16b8bf72feaf 100644 --- a/Documentation/networking/ip-sysctl.rst +++ b/Documentation/networking/ip-sysctl.rst @@ -989,14 +989,6 @@ tcp_challenge_ack_limit - INTEGER in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks) Default: 1000 -tcp_rx_skb_cache - BOOLEAN - Controls a per TCP socket cache of one skb, that might help - performance of some workloads. This might be dangerous - on systems with a lot of TCP sockets, since it increases - memory usage. - - Default: 0 (disabled) - UDP variables ============= -- cgit v1.2.3-59-g8ed1b