Re: [PATCH AUTOSEL 4.19 010/209] tcp: up initial rmem to 128KB and SYN rwin to around 64KB

On Tue, Nov 12, 2019 at 9:30 PM Sasha Levin sashal@kernel.org wrote:

From: Yuchung Cheng ycheng@google.com

[ Upstream commit a337531b942bd8a03e7052444d7e36972aac2d92 ]

Previously TCP initial receive buffer is ~87KB by default and the initial receive window is ~29KB (20 MSS). This patch changes the two numbers to 128KB and ~64KB (rounding down to the multiples of MSS) respectively. The patch also simplifies the calculations s.t. the two numbers are directly controlled by sysctl tcp_rmem[1]:

1. Initial receiver buffer budget (sk_rcvbuf): while this should be configured via sysctl tcp_rmem[1], previously tcp_fixup_rcvbuf() always override and set a larger size when a new connection establishes.

2. Initial receive window in SYN: previously it is set to 20 packets if MSS <= 1460. The number 20 was based on the initial congestion window of 10: the receiver needs twice amount to avoid being limited by the receive window upon out-of-order delivery in the first window burst. But since this only applies if the receiving MSS <= 1460, connection using large MTU (e.g. to utilize receiver zero-copy) may be limited by the receive window.

With this patch TCP memory configuration is more straight-forward and more properly sized to modern high-speed networks by default. Several popular stacks have been announcing 64KB rwin in SYNs as well.

Signed-off-by: Yuchung Cheng ycheng@google.com Signed-off-by: Wei Wang weiwan@google.com Signed-off-by: Neal Cardwell ncardwell@google.com Signed-off-by: Eric Dumazet edumazet@google.com Reviewed-by: Soheil Hassas Yeganeh soheil@google.com Signed-off-by: David S. Miller davem@davemloft.net Signed-off-by: Sasha Levin sashal@kernel.org

---

net/ipv4/tcp.c | 4 ++-- net/ipv4/tcp_input.c | 25 ++----------------------- net/ipv4/tcp_output.c | 25 ++++--------------------- 3 files changed, 8 insertions(+), 46 deletions(-)

diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 611ba174265c8..1a1fcb32c4917 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -3910,8 +3910,8 @@ void __init tcp_init(void) init_net.ipv4.sysctl_tcp_wmem[2] = max(64*1024, max_wshare);

    init_net.ipv4.sysctl_tcp_rmem[0] = SK_MEM_QUANTUM;

•   init_net.ipv4.sysctl_tcp_rmem[1] = 87380;
  

•   init_net.ipv4.sysctl_tcp_rmem[2] = max(87380, max_rshare);
  

•   init_net.ipv4.sysctl_tcp_rmem[1] = 131072;
  

•   init_net.ipv4.sysctl_tcp_rmem[2] = max(131072, max_rshare);
  
    pr_info("Hash tables configured (established %u bind %u)\n",
            tcp_hashinfo.ehash_mask + 1, tcp_hashinfo.bhash_size);
  

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 14a6a489937c1..0e2b07be08585 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -426,26 +426,7 @@ static void tcp_grow_window(struct sock *sk, const struct sk_buff *skb) } }

-/* 3. Tuning rcvbuf, when connection enters established state. */ -static void tcp_fixup_rcvbuf(struct sock *sk) -{

•   u32 mss = tcp_sk(sk)->advmss;
  

•   int rcvmem;
  

•   rcvmem = 2 * SKB_TRUESIZE(mss + MAX_TCP_HEADER) *
  

•            tcp_default_init_rwnd(mss);
  

•   /* Dynamic Right Sizing (DRS) has 2 to 3 RTT latency
  

•    * Allow enough cushion so that sender is not limited by our window
  

•    */
  

•   if (sock_net(sk)->ipv4.sysctl_tcp_moderate_rcvbuf)
  

•           rcvmem <<= 2;
  

•   if (sk->sk_rcvbuf < rcvmem)
  

•           sk->sk_rcvbuf = min(rcvmem,
  

sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); -}

-/* 4. Try to fixup all. It is made immediately after connection enters +/* 3. Try to fixup all. It is made immediately after connection enters

• established state.

*/ void tcp_init_buffer_space(struct sock *sk) @@ -454,8 +435,6 @@ void tcp_init_buffer_space(struct sock *sk) struct tcp_sock *tp = tcp_sk(sk); int maxwin;

•   if (!(sk->sk_userlocks & SOCK_RCVBUF_LOCK))
  

•           tcp_fixup_rcvbuf(sk);
    if (!(sk->sk_userlocks & SOCK_SNDBUF_LOCK))
            tcp_sndbuf_expand(sk);
  
  

@@ -485,7 +464,7 @@ void tcp_init_buffer_space(struct sock *sk) tp->snd_cwnd_stamp = tcp_jiffies32; }

-/* 5. Recalculate window clamp after socket hit its memory bounds. */ +/* 4. Recalculate window clamp after socket hit its memory bounds. */ static void tcp_clamp_window(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 2697e4397e46c..53f910bb55087 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -179,21 +179,6 @@ static inline void tcp_event_ack_sent(struct sock *sk, unsigned int pkts, inet_csk_clear_xmit_timer(sk, ICSK_TIME_DACK); }

-u32 tcp_default_init_rwnd(u32 mss) -{

•   /* Initial receive window should be twice of TCP_INIT_CWND to
  

•    * enable proper sending of new unsent data during fast recovery
  

•    * (RFC 3517, Section 4, NextSeg() rule (2)). Further place a
  

•    * limit when mss is larger than 1460.
  

•    */
  

•   u32 init_rwnd = TCP_INIT_CWND * 2;
  

•   if (mss > 1460)
  

•           init_rwnd = max((1460 * init_rwnd) / mss, 2U);
  

•   return init_rwnd;
  

-}

/* Determine a window scaling and initial window to offer.

• Based on the assumption that the given amount of space
• will be offered. Store the results in the tp structure.

@@ -228,7 +213,10 @@ void tcp_select_initial_window(const struct sock *sk, int __space, __u32 mss, if (sock_net(sk)->ipv4.sysctl_tcp_workaround_signed_windows) (*rcv_wnd) = min(space, MAX_TCP_WINDOW); else

•           (*rcv_wnd) = space;
  

•           (*rcv_wnd) = min_t(u32, space, U16_MAX);
  

•   if (init_rcv_wnd)
  

•           *rcv_wnd = min(*rcv_wnd, init_rcv_wnd * mss);
  
    (*rcv_wscale) = 0;
    if (wscale_ok) {
  

@@ -241,11 +229,6 @@ void tcp_select_initial_window(const struct sock *sk, int __space, __u32 mss, (*rcv_wscale)++; } }

•   if (!init_rcv_wnd) /* Use default unless specified otherwise */
  

•           init_rcv_wnd = tcp_default_init_rwnd(mss);
  

•   *rcv_wnd = min(*rcv_wnd, init_rcv_wnd * mss);
  

•   /* Set the clamp no higher than max representable value */
    (*window_clamp) = min_t(__u32, U16_MAX << (*rcv_wscale), *window_clamp);
  

}

2.20.1

Hi Sasha,

Apologies for not bringing this up during the review period, I only just noticed this patch when updating to v4.19.89. Can you please clarify why this patch was selected for a LTS update? This doesn't look like a bug fix and it makes significant changes to core tcp settings. The patch is also fairly old (Sep 2018) and we generally don't expect a change like this from a stable kernel update.

A follow up patch (tcp: start receiver buffer autotuning sooner) was also included in v4.19.86.

Both of the patches were selected by the auto selector based on the AUTOSEL tag. It looks like this tag only shows up in the email, have you thought about placing the tag in commit messages as well so that we can easily identify AUTOSEL patches?

Thanks, Vishwanath