Commit 31770e34 authored by Florian Westphal's avatar Florian Westphal Committed by David S. Miller
Browse files

tcp: Revert "tcp: remove header prediction" 



This reverts commit 45f119bf.

Eric Dumazet says:
  We found at Google a significant regression caused by
  45f119bf tcp: remove header prediction

  In typical RPC  (TCP_RR), when a TCP socket receives data, we now call
  tcp_ack() while we used to not call it.

  This touches enough cache lines to cause a slowdown.

so problem does not seem to be HP removal itself but the tcp_ack()
call.  Therefore, it might be possible to remove HP after all, provided
one finds a way to elide tcp_ack for most cases.

Reported-by: default avatarEric Dumazet <edumazet@google.com>
Signed-off-by: default avatarFlorian Westphal <fw@strlen.de>
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parent c1d2b4c3

include/linux/tcp.h

+6 −0
Original line number Diff line number Diff line
@@ -147,6 +147,12 @@ struct tcp_sock { 
	u16	tcp_header_len;	/* Bytes of tcp header to send		*/

	u16	gso_segs;	/* Max number of segs per GSO packet	*/



/*

 *	Header prediction flags

 *	0x5?10 << 16 + snd_wnd in net byte order

 */

	__be32	pred_flags;



/*

 *	RFC793 variables by their proper names. This means you can

 *	read the code and the spec side by side (and laugh ...)

include/net/tcp.h

+23 −0
Original line number Diff line number Diff line
@@ -634,6 +634,29 @@ static inline u32 __tcp_set_rto(const struct tcp_sock *tp) 
	return usecs_to_jiffies((tp->srtt_us >> 3) + tp->rttvar_us);

}



static inline void __tcp_fast_path_on(struct tcp_sock *tp, u32 snd_wnd)

{

	tp->pred_flags = htonl((tp->tcp_header_len << 26) |

			       ntohl(TCP_FLAG_ACK) |

			       snd_wnd);

}



static inline void tcp_fast_path_on(struct tcp_sock *tp)

{

	__tcp_fast_path_on(tp, tp->snd_wnd >> tp->rx_opt.snd_wscale);

}



static inline void tcp_fast_path_check(struct sock *sk)

{

	struct tcp_sock *tp = tcp_sk(sk);



	if (RB_EMPTY_ROOT(&tp->out_of_order_queue) &&

	    tp->rcv_wnd &&

	    atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf &&

	    !tp->urg_data)

		tcp_fast_path_on(tp);

}



/* Compute the actual rto_min value */

static inline u32 tcp_rto_min(struct sock *sk)

{

include/uapi/linux/snmp.h

+2 −0
Original line number Diff line number Diff line
@@ -184,7 +184,9 @@ enum 
	LINUX_MIB_DELAYEDACKLOST,		/* DelayedACKLost */

	LINUX_MIB_LISTENOVERFLOWS,		/* ListenOverflows */

	LINUX_MIB_LISTENDROPS,			/* ListenDrops */

	LINUX_MIB_TCPHPHITS,			/* TCPHPHits */

	LINUX_MIB_TCPPUREACKS,			/* TCPPureAcks */

	LINUX_MIB_TCPHPACKS,			/* TCPHPAcks */

	LINUX_MIB_TCPRENORECOVERY,		/* TCPRenoRecovery */

	LINUX_MIB_TCPSACKRECOVERY,		/* TCPSackRecovery */

	LINUX_MIB_TCPSACKRENEGING,		/* TCPSACKReneging */

net/ipv4/proc.c

+2 −0
Original line number Diff line number Diff line
@@ -206,7 +206,9 @@ static const struct snmp_mib snmp4_net_list[] = { 
	SNMP_MIB_ITEM("DelayedACKLost", LINUX_MIB_DELAYEDACKLOST),

	SNMP_MIB_ITEM("ListenOverflows", LINUX_MIB_LISTENOVERFLOWS),

	SNMP_MIB_ITEM("ListenDrops", LINUX_MIB_LISTENDROPS),

	SNMP_MIB_ITEM("TCPHPHits", LINUX_MIB_TCPHPHITS),

	SNMP_MIB_ITEM("TCPPureAcks", LINUX_MIB_TCPPUREACKS),

	SNMP_MIB_ITEM("TCPHPAcks", LINUX_MIB_TCPHPACKS),

	SNMP_MIB_ITEM("TCPRenoRecovery", LINUX_MIB_TCPRENORECOVERY),

	SNMP_MIB_ITEM("TCPSackRecovery", LINUX_MIB_TCPSACKRECOVERY),

	SNMP_MIB_ITEM("TCPSACKReneging", LINUX_MIB_TCPSACKRENEGING),

net/ipv4/tcp.c

+3 −1
Original line number Diff line number Diff line
@@ -1963,8 +1963,10 @@ int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock, 
		tcp_rcv_space_adjust(sk);



skip_copy:

		if (tp->urg_data && after(tp->copied_seq, tp->urg_seq))

		if (tp->urg_data && after(tp->copied_seq, tp->urg_seq)) {

			tp->urg_data = 0;

			tcp_fast_path_check(sk);

		}

		if (used + offset < skb->len)

			continue;

CRA Git | Maintained and supported by SUSTech CRA and CCSE