| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <1376940568-16512-2-git-send-email-christoph.paasch@uclouvain.be>
Date: Mon, 19 Aug 2013 21:29:27 +0200
From: Christoph Paasch <christoph.paasch@...ouvain.be>
To: Eric Dumazet <eric.dumazet@...il.com>
Cc: netdev@...r.kernel.org, Phil Oester <kernel@...uxace.com>,
Corey Hickey <bugfood-ml@...ooh.org>,
Benjamin Hesmans <benjamin.hesmans@...ouvain.be>
Subject: [RFC 1/2] Use acked_out for reno-style ack acounting instead of sacked_out
This patch adds acked_out to tcp_sock, so that all acounting of acked
packets in case of non-sack connections is done within this field.
tcp_acked_out() sums up sacked_out and acked_out.
Tested-by: Benjamin Hesmans <benjamin.hesmans@...ouvain.be>
Signed-off-by: Christoph Paasch <christoph.paasch@...ouvain.be>
---
include/linux/tcp.h | 1 +
include/net/tcp.h | 17 ++++++++++++-----
net/ipv4/tcp_input.c | 44 +++++++++++++++++++++++++++-----------------
net/ipv4/tcp_minisocks.c | 1 +
net/ipv4/tcp_output.c | 2 +-
net/ipv4/tcp_timer.c | 2 +-
6 files changed, 43 insertions(+), 24 deletions(-)
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index d686334..6622595 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -241,6 +241,7 @@ struct tcp_sock {
u32 pushed_seq; /* Last pushed seq, required to talk to windows */
u32 lost_out; /* Lost packets */
u32 sacked_out; /* SACK'd packets */
+ u32 acked_out; /* DupACK'd packets */
u32 fackets_out; /* FACK'd packets */
u32 tso_deferred;
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 09cb5c1..8210057 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -871,9 +871,14 @@ static inline void tcp_disable_early_retrans(struct tcp_sock *tp)
tp->do_early_retrans = 0;
}
+static inline unsigned int tcp_acked_out(const struct tcp_sock *tp)
+{
+ return tp->sacked_out + tp->acked_out;
+}
+
static inline unsigned int tcp_left_out(const struct tcp_sock *tp)
{
- return tp->sacked_out + tp->lost_out;
+ return tcp_acked_out(tp) + tp->lost_out;
}
/* This determines how many packets are "in the network" to the best
@@ -1446,12 +1451,12 @@ static inline void tcp_push_pending_frames(struct sock *sk)
}
/* Start sequence of the skb just after the highest skb with SACKed
- * bit, valid only if sacked_out > 0 or when the caller has ensured
+ * bit, valid only if sacked_out or acked_out > 0 or when the caller has ensured
* validity by itself.
*/
static inline u32 tcp_highest_sack_seq(struct tcp_sock *tp)
{
- if (!tp->sacked_out)
+ if (!tp->sacked_out && !tp->acked_out)
return tp->snd_una;
if (tp->highest_sack == NULL)
@@ -1481,8 +1486,10 @@ static inline void tcp_highest_sack_combine(struct sock *sk,
struct sk_buff *old,
struct sk_buff *new)
{
- if (tcp_sk(sk)->sacked_out && (old == tcp_sk(sk)->highest_sack))
- tcp_sk(sk)->highest_sack = new;
+ struct tcp_sock *tp = tcp_sk(sk);
+
+ if ((tp->sacked_out || tp->acked_out) && (old == tp->highest_sack))
+ tp->highest_sack = new;
}
/* Determines whether this is a thin stream (which may suffer from
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index e965cc7..af764d0 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -1748,8 +1748,8 @@ out:
return state.flag;
}
-/* Limits sacked_out so that sum with lost_out isn't ever larger than
- * packets_out. Returns false if sacked_out adjustement wasn't necessary.
+/* Limits acked_out so that sum with lost_out isn't ever larger than
+ * packets_out. Returns false if acked_out adjustement wasn't necessary.
*/
static bool tcp_limit_reno_sacked(struct tcp_sock *tp)
{
@@ -1758,8 +1758,8 @@ static bool tcp_limit_reno_sacked(struct tcp_sock *tp)
holes = max(tp->lost_out, 1U);
holes = min(holes, tp->packets_out);
- if ((tp->sacked_out + holes) > tp->packets_out) {
- tp->sacked_out = tp->packets_out - holes;
+ if ((tp->acked_out + holes) > tp->packets_out) {
+ tp->acked_out = tp->packets_out - holes;
return true;
}
return false;
@@ -1781,7 +1781,7 @@ static void tcp_check_reno_reordering(struct sock *sk, const int addend)
static void tcp_add_reno_sack(struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
- tp->sacked_out++;
+ tp->acked_out++;
tcp_check_reno_reordering(sk, 0);
tcp_verify_left_out(tp);
}
@@ -1794,10 +1794,10 @@ static void tcp_remove_reno_sacks(struct sock *sk, int acked)
if (acked > 0) {
/* One ACK acked hole. The rest eat duplicate ACKs. */
- if (acked - 1 >= tp->sacked_out)
- tp->sacked_out = 0;
+ if (acked - 1 >= tp->acked_out)
+ tp->acked_out = 0;
else
- tp->sacked_out -= acked - 1;
+ tp->acked_out -= acked - 1;
}
tcp_check_reno_reordering(sk, acked);
tcp_verify_left_out(tp);
@@ -1805,7 +1805,7 @@ static void tcp_remove_reno_sacks(struct sock *sk, int acked)
static inline void tcp_reset_reno_sack(struct tcp_sock *tp)
{
- tp->sacked_out = 0;
+ tp->acked_out = 0;
}
static void tcp_clear_retrans_partial(struct tcp_sock *tp)
@@ -1823,6 +1823,7 @@ void tcp_clear_retrans(struct tcp_sock *tp)
tp->fackets_out = 0;
tp->sacked_out = 0;
+ tp->acked_out = 0;
}
/* Enter Loss state. If "how" is not zero, forget all SACK information
@@ -1857,6 +1858,7 @@ void tcp_enter_loss(struct sock *sk, int how)
tp->undo_marker = tp->snd_una;
if (how) {
tp->sacked_out = 0;
+ tp->acked_out = 0;
tp->fackets_out = 0;
}
tcp_clear_all_retrans_hints(tp);
@@ -1921,7 +1923,7 @@ static bool tcp_check_sack_reneging(struct sock *sk, int flag)
static inline int tcp_fackets_out(const struct tcp_sock *tp)
{
- return tcp_is_reno(tp) ? tp->sacked_out + 1 : tp->fackets_out;
+ return tcp_is_reno(tp) ? tp->acked_out + 1 : tp->fackets_out;
}
/* Heurestics to calculate number of duplicate ACKs. There's no dupACKs
@@ -1941,7 +1943,7 @@ static inline int tcp_fackets_out(const struct tcp_sock *tp)
*/
static inline int tcp_dupack_heuristics(const struct tcp_sock *tp)
{
- return tcp_is_fack(tp) ? tp->fackets_out : tp->sacked_out + 1;
+ return tcp_is_fack(tp) ? tp->fackets_out : tcp_acked_out(tp) + 1;
}
static bool tcp_pause_early_retransmit(struct sock *sk, int flag)
@@ -2077,7 +2079,7 @@ static bool tcp_time_to_recover(struct sock *sk, int flag)
*/
packets_out = tp->packets_out;
if (packets_out <= tp->reordering &&
- tp->sacked_out >= max_t(__u32, packets_out/2, sysctl_tcp_reordering) &&
+ tcp_acked_out(tp) >= max_t(__u32, packets_out/2, sysctl_tcp_reordering) &&
!tcp_may_send_now(sk)) {
/* We have nothing to send. This connection is limited
* either by receiver window or by application.
@@ -2100,8 +2102,8 @@ static bool tcp_time_to_recover(struct sock *sk, int flag)
* Mitigation A.3 in the RFC and delay the retransmission for a short
* interval if appropriate.
*/
- if (tp->do_early_retrans && !tp->retrans_out && tp->sacked_out &&
- (tp->packets_out >= (tp->sacked_out + 1) && tp->packets_out < 4) &&
+ if (tp->do_early_retrans && !tp->retrans_out && (tp->sacked_out || tp->acked_out) &&
+ (tp->packets_out >= (tcp_acked_out(tp) + 1) && tp->packets_out < 4) &&
!tcp_may_send_now(sk))
return !tcp_pause_early_retransmit(sk, flag);
@@ -2701,9 +2703,11 @@ static void tcp_fastretrans_alert(struct sock *sk, const int acked,
(tcp_fackets_out(tp) > tp->reordering));
int fast_rexmit = 0;
- if (WARN_ON(!tp->packets_out && tp->sacked_out))
+ if (WARN_ON(!tp->packets_out && (tp->sacked_out || tp->acked_out))) {
tp->sacked_out = 0;
- if (WARN_ON(!tp->sacked_out && tp->fackets_out))
+ tp->acked_out = 0;
+ }
+ if (WARN_ON(!tp->sacked_out && !tp->acked_out && tp->fackets_out))
tp->fackets_out = 0;
/* Now state machine starts.
@@ -3078,6 +3082,7 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets,
#if FASTRETRANS_DEBUG > 0
WARN_ON((int)tp->sacked_out < 0);
+ WARN_ON((int)tp->acked_out < 0);
WARN_ON((int)tp->lost_out < 0);
WARN_ON((int)tp->retrans_out < 0);
if (!tp->packets_out && tcp_is_sack(tp)) {
@@ -3092,6 +3097,11 @@ static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets,
tp->sacked_out, icsk->icsk_ca_state);
tp->sacked_out = 0;
}
+ if (tp->acked_out) {
+ pr_debug("Leak a=%u %d\n",
+ tp->acked_out, icsk->icsk_ca_state);
+ tp->acked_out = 0;
+ }
if (tp->retrans_out) {
pr_debug("Leak r=%u %d\n",
tp->retrans_out, icsk->icsk_ca_state);
@@ -3270,7 +3280,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag)
u32 prior_in_flight;
u32 prior_fackets;
int prior_packets = tp->packets_out;
- const int prior_unsacked = tp->packets_out - tp->sacked_out;
+ const int prior_unsacked = tp->packets_out - tcp_acked_out(tp);
int acked = 0; /* Number of packets newly acked */
s32 sack_rtt = -1;
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 58a3e69..a2f5279 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -407,6 +407,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
newtp->packets_out = 0;
newtp->retrans_out = 0;
newtp->sacked_out = 0;
+ newtp->acked_out = 0;
newtp->fackets_out = 0;
newtp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
tcp_enable_early_retrans(newtp);
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 884efff..17ebbff 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1033,7 +1033,7 @@ static void tcp_adjust_pcount(struct sock *sk, const struct sk_buff *skb, int de
/* Reno case is special. Sigh... */
if (tcp_is_reno(tp) && decr > 0)
- tp->sacked_out -= min_t(u32, tp->sacked_out, decr);
+ tp->acked_out -= min_t(u32, tp->acked_out, decr);
tcp_adjust_fackets_out(sk, skb, decr);
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 4b85e6f..82acaea 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -405,7 +405,7 @@ void tcp_retransmit_timer(struct sock *sk)
} else if (icsk->icsk_ca_state == TCP_CA_Loss) {
mib_idx = LINUX_MIB_TCPLOSSFAILURES;
} else if ((icsk->icsk_ca_state == TCP_CA_Disorder) ||
- tp->sacked_out) {
+ tp->sacked_out || tp->acked_out) {
if (tcp_is_sack(tp))
mib_idx = LINUX_MIB_TCPSACKFAILURES;
else
--
1.8.1.2
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists