Several spots in the kernel perform a sequence like:
skb_queue_tail(&sk->s_receive_queue, skb);
sk->sk_data_ready(sk, skb->len);
But at the moment we place the SKB onto the socket receive queue it
can be consumed and freed up. So this skb->len access is potentially
to freed up memory.
Furthermore, the skb->len can be modified by the consumer so it is
possible that the value isn't accurate.
And finally, no actual implementation of this callback actually uses
the length argument. And since nobody actually cared about it's
value, lots of call sites pass arbitrary values in such as '0' and
even '1'.
So just remove the length argument from the callback, that way there
is no confusion whatsoever and all of these use-after-free cases get
fixed as a side effect.
Based upon a patch by Eric Dumazet and his suggestion to audit this
issue tree-wide.
Signed-off-by: David S. Miller <davem@davemloft.net>
All skb in socket write queue should be properly timestamped.
In case of FastOpen, we special case the SYN+DATA 'message' as we
queue in socket wrote queue the two fallback skbs:
1) SYN message by itself.
2) DATA segment by itself.
We should make sure these skbs have proper timestamps.
Add a WARN_ON_ONCE() to eventually catch future violations.
Fixes: 740b0f1841 ("tcp: switch rtt estimations to usec resolution")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/wireless/ath/ath9k/recv.c
drivers/net/wireless/mwifiex/pcie.c
net/ipv6/sit.c
The SIT driver conflict consists of a bug fix being done by hand
in 'net' (missing u64_stats_init()) whilst in 'net-next' a helper
was created (netdev_alloc_pcpu_stats()) which takes care of this.
The two wireless conflicts were overlapping changes.
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the following snmp stats:
TCPFastOpenActiveFail: Fast Open attempts (SYN/data) failed beacuse
the remote does not accept it or the attempts timed out.
TCPSynRetrans: number of SYN and SYN/ACK retransmits to break down
retransmissions into SYN, fast-retransmits, timeout retransmits, etc.
TCPOrigDataSent: number of outgoing packets with original data (excluding
retransmission but including data-in-SYN). This counter is different from
TcpOutSegs because TcpOutSegs also tracks pure ACKs. TCPOrigDataSent is
more useful to track the TCP retransmission rate.
Change TCPFastOpenActive to track only successful Fast Opens to be symmetric to
TCPFastOpenPassive.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: Lawrence Brakmo <brakmo@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RTT may be bogus with tall loss probe (TLP) when a packet
is retransmitted and latter (s)acked without TCPCB_SACKED_RETRANS flag.
For example, TLP calls __tcp_retransmit_skb() instead of
tcp_retransmit_skb(). The skb timestamps are updated but the sacked
flag is not marked with TCPCB_SACKED_RETRANS. As a result we'll
get bogus RTT in tcp_clean_rtx_queue() or in tcp_sacktag_one() on
spurious retransmission.
The fix is to apply the sticky flag TCP_EVER_RETRANS to enforce Karn's
check on RTT sampling. However this will disable F-RTO if timeout occurs
after TLP, by resetting undo_marker in tcp_enter_loss(). We relax this
check to only if any pending retransmists are still in-flight.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Upcoming congestion controls for TCP require usec resolution for RTT
estimations. Millisecond resolution is simply not enough these days.
FQ/pacing in DC environments also require this change for finer control
and removal of bimodal behavior due to the current hack in
tcp_update_pacing_rate() for 'small rtt'
TCP_CONG_RTT_STAMP is no longer needed.
As Julian Anastasov pointed out, we need to keep user compatibility :
tcp_metrics used to export RTT and RTTVAR in msec resolution,
so we added RTT_US and RTTVAR_US. An iproute2 patch is needed
to use the new attributes if provided by the kernel.
In this example ss command displays a srtt of 32 usecs (10Gbit link)
lpk51:~# ./ss -i dst lpk52
Netid State Recv-Q Send-Q Local Address:Port Peer
Address:Port
tcp ESTAB 0 1 10.246.11.51:42959
10.246.11.52:64614
cubic wscale:6,6 rto:201 rtt:0.032/0.001 ato:40 mss:1448
cwnd:10 send
3620.0Mbps pacing_rate 7240.0Mbps unacked:1 rcv_rtt:993 rcv_space:29559
Updated iproute2 ip command displays :
lpk51:~# ./ip tcp_metrics | grep 10.246.11.52
10.246.11.52 age 561.914sec cwnd 10 rtt 274us rttvar 213us source
10.246.11.51
Old binary displays :
lpk51:~# ip tcp_metrics | grep 10.246.11.52
10.246.11.52 age 561.914sec cwnd 10 rtt 250us rttvar 125us source
10.246.11.51
With help from Julian Anastasov, Stephen Hemminger and Yuchung Cheng
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Larry Brakmo <brakmo@google.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP pacing depends on an accurate srtt estimation.
Current srtt estimation is using jiffie resolution,
and has an artificial offset of at least 1 ms, which can produce
slowdowns when FQ/pacing is used, especially in DC world,
where typical rtt is below 1 ms.
We are planning a switch to usec resolution for linux-3.15,
but in the meantime, this patch removes the 1 ms offset.
All we need is to have tp->srtt minimal value of 1 to differentiate
the case of srtt being initialized or not, not 8.
The problematic behavior was observed on a 40Gbit testbed,
where 32 concurrent netperf were reaching 12Gbps of aggregate
speed, instead of line speed.
This patch also has the effect of reporting more accurate srtt and send
rates to iproute2 ss command as in :
$ ss -i dst cca2
Netid State Recv-Q Send-Q Local Address:Port
Peer Address:Port
tcp ESTAB 0 0 10.244.129.1:56984
10.244.129.2:12865
cubic wscale:6,6 rto:200 rtt:0.25/0.25 ato:40 mss:1448 cwnd:10 send
463.4Mbps rcv_rtt:1 rcv_space:29200
tcp ESTAB 0 390960 10.244.129.1:60247
10.244.129.2:50204
cubic wscale:6,6 rto:200 rtt:0.875/0.75 mss:1448 cwnd:73 ssthresh:51
send 966.4Mbps unacked:73 retrans:0/121 rcv_space:29200
Reported-by: Vytautas Valancius <valas@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The following are only used in one file:
tcp_connect_init
tcp_set_rto
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix checkpatch errors like:
ERROR: spaces required around that XXX
Signed-off-by: Weilong Chen <chenweilong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Slow start now increases cwnd by 1 if an ACK acknowledges some packets,
regardless the number of packets. Consequently slow start performance
is highly dependent on the degree of the stretch ACKs caused by
receiver or network ACK compression mechanisms (e.g., delayed-ACK,
GRO, etc). But slow start algorithm is to send twice the amount of
packets of packets left so it should process a stretch ACK of degree
N as if N ACKs of degree 1, then exits when cwnd exceeds ssthresh. A
follow up patch will use the remainder of the N (if greater than 1)
to adjust cwnd in the congestion avoidance phase.
In addition this patch retires the experimental limited slow start
(LSS) feature. LSS has multiple drawbacks but questionable benefit. The
fractional cwnd increase in LSS requires a loop in slow start even
though it's rarely used. Configuring such an increase step via a global
sysctl on different BDPS seems hard. Finally and most importantly the
slow start overshoot concern is now better covered by the Hybrid slow
start (hystart) enabled by default.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/emulex/benet/be.h
drivers/net/netconsole.c
net/bridge/br_private.h
Three mostly trivial conflicts.
The net/bridge/br_private.h conflict was a function signature (argument
addition) change overlapping with the extern removals from Joe Perches.
In drivers/net/netconsole.c we had one change adjusting a printk message
whilst another changed "printk(KERN_INFO" into "pr_info(".
Lastly, the emulex change was a new inline function addition overlapping
with Joe Perches's extern removals.
Signed-off-by: David S. Miller <davem@davemloft.net>
Patch ed08495c3 "tcp: use RTT from SACK for RTO" always re-arms RTO upon
obtaining a RTT sample from newly sacked data.
But technically RTO should only be re-armed when the data sent before
the last (re)transmission of write queue head are (s)acked. Otherwise
the RTO may continue to extend during loss recovery on data sent
in the future.
Note that RTTs from ACK or timestamps do not have this problem, as the RTT
source must be from data sent before.
The new RTO re-arm policy is
1) Always re-arm RTO if SND.UNA is advanced
2) Re-arm RTO if sack RTT is available, provided the sacked data was
sent before the last time write_queue_head was sent.
Signed-off-by: Larry Brakmo <brakmo@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Patch ed08495c3 "tcp: use RTT from SACK for RTO" has a bug that
it does not check if the ACK acknowledge new data before taking
the RTT sample from TCP timestamps. This patch adds the check
back as required by the RFC.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tp->lsndtime may not always be the SYNACK timestamp if a passive
Fast Open socket sends data before handshake completes. And if the
remote acknowledges both the data and the SYNACK, the RTT sample
is already taken in tcp_ack(), so no need to call
tcp_update_ack_rtt() in tcp_synack_rtt_meas() aagain.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/usb/qmi_wwan.c
include/net/dst.h
Trivial merge conflicts, both were overlapping changes.
Signed-off-by: David S. Miller <davem@davemloft.net>
For passive TCP connections, upon receiving the ACK that completes the
3WHS, make sure we set our pacing rate after we get our first RTT
sample.
On passive TCP connections, when we receive the ACK completing the
3WHS we do not take an RTT sample in tcp_ack(), but rather in
tcp_synack_rtt_meas(). So upon receiving the ACK that completes the
3WHS, tcp_ack() leaves sk_pacing_rate at its initial value.
Originally the initial sk_pacing_rate value was 0, so passive-side
connections defaulted to sysctl_tcp_min_tso_segs (2 segs) in skbuffs
made in the first RTT. With a default initial cwnd of 10 packets, this
happened to be correct for RTTs 5ms or bigger, so it was hard to
see problems in WAN or emulated WAN testing.
Since 7eec4174ff ("pkt_sched: fq: fix non TCP flows pacing"), the
initial sk_pacing_rate is 0xffffffff. So after that change, passive
TCP connections were keeping this value (and using large numbers of
segments per skbuff) until receiving an ACK for data.
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On receiving an ACK that covers the loss probe sequence, TLP
immediately sets the congestion state to Open, even though some packets
are not recovered and retransmisssion are on the way. The later ACks
may trigger a WARN_ON check in step D of tcp_fastretrans_alert(), e.g.,
https://bugzilla.redhat.com/show_bug.cgi?id=989251
The fix is to follow the similar procedure in recovery by calling
tcp_try_keep_open(). The sender switches to Open state if no packets
are retransmissted. Otherwise it goes to Disorder and let subsequent
ACKs move the state to Recovery or Open.
Reported-By: Michael Sterrett <michael@sterretts.net>
Tested-By: Dormando <dormando@rydia.net>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk_pacing_rate is read by sch_fq packet scheduler at any time,
with no synchronization, so make sure we update it in a
sensible way. ACCESS_ONCE() is how we instruct compiler
to not do stupid things, like using the memory location
as a temporary variable.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
include/linux/netdevice.h
net/core/sock.c
Trivial merge issues.
Removal of "extern" for functions declaration in netdevice.h
at the same time "const" was added to an argument.
Two parallel line additions in net/core/sock.c
Signed-off-by: David S. Miller <davem@davemloft.net>
Yuchung found following problem :
There are bugs in the SACK processing code, merging part in
tcp_shift_skb_data(), that incorrectly resets or ignores the sacked
skbs FIN flag. When a receiver first SACK the FIN sequence, and later
throw away ofo queue (e.g., sack-reneging), the sender will stop
retransmitting the FIN flag, and hangs forever.
Following packetdrill test can be used to reproduce the bug.
$ cat sack-merge-bug.pkt
`sysctl -q net.ipv4.tcp_fack=0`
// Establish a connection and send 10 MSS.
0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+.000 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+.000 bind(3, ..., ...) = 0
+.000 listen(3, 1) = 0
+.050 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop,nop,wscale 7>
+.000 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 6>
+.001 < . 1:1(0) ack 1 win 1024
+.000 accept(3, ..., ...) = 4
+.100 write(4, ..., 12000) = 12000
+.000 shutdown(4, SHUT_WR) = 0
+.000 > . 1:10001(10000) ack 1
+.050 < . 1:1(0) ack 2001 win 257
+.000 > FP. 10001:12001(2000) ack 1
+.050 < . 1:1(0) ack 2001 win 257 <sack 10001:11001,nop,nop>
+.050 < . 1:1(0) ack 2001 win 257 <sack 10001:12002,nop,nop>
// SACK reneg
+.050 < . 1:1(0) ack 12001 win 257
+0 %{ print "unacked: ",tcpi_unacked }%
+5 %{ print "" }%
First, a typo inverted left/right of one OR operation, then
code forgot to advance end_seq if the merged skb carried FIN.
Bug was added in 2.6.29 by commit 832d11c5cd
("tcp: Try to restore large SKBs while SACK processing")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Cc: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
The variable fully_acked is only assigned the values true and false.
Change its type to bool.
The simplified semantic patch that find this problem is as
follows (http://coccinelle.lip6.fr/):
@exists@
type T;
identifier b;
@@
- T
+ bool
b = ...;
... when any
b = \(true\|false\)
Signed-off-by: Peter Senna Tschudin <peter.senna@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_fixup_sndbuf() is underestimating initial send buffer requirements.
It was not noticed because big GSO packets were escaping the limitation,
but with smaller TSO packets (or TSO/GSO/SG off), application hits
sk_sndbuf before having a chance to fill enough packets in socket write
queue.
- initial cwnd can be bigger than 10 for specific routes
- SKB_TRUESIZE() is a bit under real needs in some cases,
because of power-of-two rounding in kmalloc()
- Fast Recovery (RFC 5681 3.2) : Cubic needs 70% factor
- Extra cushion (application might react slowly to POLLOUT)
tcp_v4_conn_req_fastopen() needs to call tcp_init_metrics() before
calling tcp_init_buffer_space()
Then we realize tcp_new_space() should call tcp_fixup_sndbuf()
instead of duplicating this stuff.
Rename tcp_fixup_sndbuf() to tcp_sndbuf_expand() to be more
descriptive.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As mentioned in commit afe4fd0624 ("pkt_sched: fq: Fair Queue packet
scheduler"), this patch adds a new socket option.
SO_MAX_PACING_RATE offers the application the ability to cap the
rate computed by transport layer. Value is in bytes per second.
u32 val = 1000000;
setsockopt(sockfd, SOL_SOCKET, SO_MAX_PACING_RATE, &val, sizeof(val));
To be effectively paced, a flow must use FQ packet scheduler.
Note that a packet scheduler takes into account the headers for its
computations. The effective payload rate depends on MSS and retransmits
if any.
I chose to make this pacing rate a SOL_SOCKET option instead of a
TCP one because this can be used by other protocols.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Steinar H. Gunderson <sesse@google.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Dynamic Right Sizing (DRS) is supposed to open TCP receive window
automatically, but suffers from two bugs, presented by order
of importance.
1) tcp_rcv_space_adjust() fix :
Using twice the last received amount is very pessimistic,
because it doesn't allow fast recovery or proper slow start
ramp up, if sender wants to increase cwin by 100% every RTT.
copied = bytes received in previous RTT
2*copied = bytes we expect to receive in next RTT
4*copied = bytes we need to advertise in rwin at end of next RTT
DRS is one RTT late, it needs a 4x factor.
If sender is not using ABC, and increases cwin by 50% every rtt,
then we needed 1.5*1.5 = 2.25 factor.
This is probably why this bug was not really noticed.
2) There is no window adjustment after first RTT. DRS triggers only
after the second RTT.
DRS needs two RTT to initialize, so tcp_fixup_rcvbuf() should setup
sk_rcvbuf to allow proper window grow for first two RTT.
This patch increases TCP efficiency particularly for large RTT flows
when autotuning is used at the receiver, and more particularly
in presence of packet losses.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Cc: Van Jacobson <vanj@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP receive window handling is multi staged.
A socket has a memory budget, static or dynamic, in sk_rcvbuf.
Because we do not really know how this memory budget translates to
a TCP window (payload), TCP announces a small initial window
(about 20 MSS).
When a packet is received, we increase TCP rcv_win depending
on the payload/truesize ratio of this packet. Good citizen
packets give a hint that it's reasonable to have rcv_win = sk_rcvbuf/2
This heuristic takes place in tcp_grow_window()
Problem is : We currently call tcp_grow_window() only for in-order
packets.
This means that reorders or packet losses stop proper grow of
rcv_win, and senders are unable to benefit from fast recovery,
or proper reordering level detection.
Really, a packet being stored in OFO queue is not a bad citizen.
It should be part of the game as in-order packets.
In our traces, we very often see sender is limited by linux small
receive windows, even if linux hosts use autotuning (DRS) and should
allow rcv_win to grow to ~3MB.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 0f7cc9a3 "tcp: increase throughput when reordering is high",
it only allows cwnd to increase in Open state. This mistakenly disables
slow start after timeout (CA_Loss). Moreover cwnd won't grow if the
state moves from Disorder to Open later in tcp_fastretrans_alert().
Therefore the correct logic should be to allow cwnd to grow as long
as the data is received in order in Open, Loss, or even Disorder state.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
net/bridge/br_multicast.c
net/ipv6/sit.c
The conflicts were minor:
1) sit.c changes overlap with change to ip_tunnel_xmit() signature.
2) br_multicast.c had an overlap between computing max_delay using
msecs_to_jiffies and turning MLDV2_MRC() into an inline function
with a name using lowercase instead of uppercase letters.
3) stmmac had two overlapping changes, one which conditionally allocated
and hooked up a dma_cfg based upon the presence of the pbl OF property,
and another one handling store-and-forward DMA made. The latter of
which should not go into the new of_find_property() basic block.
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_rcv_established() returns only one value namely 0. We change the return
value to void (as suggested by David Miller).
After commit 0c24604b (tcp: implement RFC 5961 4.2), we no longer send RSTs in
response to SYNs. We can remove the check and processing on the return value of
tcp_rcv_established().
We also fix jtcp_rcv_established() in tcp_probe.c to match that of
tcp_rcv_established().
Signed-off-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After hearing many people over past years complaining against TSO being
bursty or even buggy, we are proud to present automatic sizing of TSO
packets.
One part of the problem is that tcp_tso_should_defer() uses an heuristic
relying on upcoming ACKS instead of a timer, but more generally, having
big TSO packets makes little sense for low rates, as it tends to create
micro bursts on the network, and general consensus is to reduce the
buffering amount.
This patch introduces a per socket sk_pacing_rate, that approximates
the current sending rate, and allows us to size the TSO packets so
that we try to send one packet every ms.
This field could be set by other transports.
Patch has no impact for high speed flows, where having large TSO packets
makes sense to reach line rate.
For other flows, this helps better packet scheduling and ACK clocking.
This patch increases performance of TCP flows in lossy environments.
A new sysctl (tcp_min_tso_segs) is added, to specify the
minimal size of a TSO packet (default being 2).
A follow-up patch will provide a new packet scheduler (FQ), using
sk_pacing_rate as an input to perform optional per flow pacing.
This explains why we chose to set sk_pacing_rate to twice the current
rate, allowing 'slow start' ramp up.
sk_pacing_rate = 2 * cwnd * mss / srtt
v2: Neal Cardwell reported a suspect deferring of last two segments on
initial write of 10 MSS, I had to change tcp_tso_should_defer() to take
into account tp->xmit_size_goal_segs
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Van Jacobson <vanj@google.com>
Cc: Tom Herbert <therbert@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The zero value means that tsecr is not valid, so it's a special case.
tsoffset is used to customize tcp_time_stamp for one socket.
tsoffset is usually zero, it's used when a socket was moved from one
host to another host.
Currently this issue affects logic of tcp_rcv_rtt_measure_ts. Due to
incorrect value of rcv_tsecr, tcp_rcv_rtt_measure_ts sets rto to
TCP_RTO_MAX.
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: James Morris <jmorris@namei.org>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Patrick McHardy <kaber@trash.net>
Reported-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: Andrey Vagin <avagin@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The stack currently detects reordering and avoid spurious
retransmission very well. However the throughput is sub-optimal under
high reordering because cwnd is increased only if the data is deliverd
in order. I.e., FLAG_DATA_ACKED check in tcp_ack(). The more packet
are reordered the worse the throughput is.
Therefore when reordering is proven high, cwnd should advance whenever
the data is delivered regardless of its ordering. If reordering is low,
conservatively advance cwnd only on ordered deliveries in Open state,
and retain cwnd in Disordered state (RFC5681).
Using netperf on a qdisc setup of 20Mbps BW and random RTT from 45ms
to 55ms (for reordering effect). This change increases TCP throughput
by 20 - 25% to near bottleneck BW.
A special case is the stretched ACK with new SACK and/or ECE mark.
For example, a receiver may receive an out of order or ECN packet with
unacked data buffered because of LRO or delayed ACK. The principle on
such an ACK is to advance cwnd on the cummulative acked part first,
then reduce cwnd in tcp_fastretrans_alert().
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
On timeout the TCP sender unconditionally resets the estimated degree
of network reordering (tp->reordering). The idea behind this is that
the estimate is too large to trigger fast recovery (e.g., due to a IP
path change).
But for example if the sender only had 2 packets outstanding, then a
timeout doesn't tell much about reordering. A sender that learns about
reordering on big writes and loses packets on small writes will end up
falsely retransmitting again and again, especially when reordering is
more likely on big writes.
Therefore the sender should only suspect that tp->reordering is too
high if it could have gone into fast recovery with the (lower) default
estimate.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If RTT is not available because Karn's check has failed or no
new packet is acked, use the RTT measured from SACK to estimate
the RTO. The sender can continue to estimate the RTO during loss
recovery or reordering event upon receiving non-partial ACKs.
This also changes when the RTO is re-armed. Previously it is
only re-armed when some data is cummulatively acknowledged (i.e.,
SND.UNA advances), but now it is re-armed whenever RTT estimator
is updated. This feature is particularly useful to reduce spurious
timeout for buffer bloat including cellular carriers [1], and
RTT estimation on reordering events.
[1] "An In-depth Study of LTE: Effect of Network Protocol and
Application Behavior on Performance", In Proc. of SIGCOMM 2013
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Take RTT sample if an ACK selectively acks some sequences that
have never been retransmitted. The Karn's algorithm does not apply
even if that ACK (s)acks other retransmitted sequences, because it
must been generated by an original but perhaps out-of-order packet.
There is no ambiguity. In case when multiple blocks are newly
sacked because of ACK losses the earliest block is used to
measure RTT, similar to cummulative ACKs.
Such RTT samples allow the sender to estimate the RTO during loss
recovery and packet reordering events. It is still useful even with
TCP timestamps. That's because during these events the SND.UNA may
not advance preventing RTT samples from TS ECR (thus the FLAG_ACKED
check before calling tcp_ack_update_rtt()). Therefore this new
RTT source is complementary to existing ACK and TS RTT mechanisms.
This patch does not update the RTO. It is done in the next patch.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Prefer packet timings to TS-ecr for RTT measurements when both
sources are available. That's because broken middle-boxes and remote
peer can return packets with corrupted TS ECR fields. Similarly most
congestion controls that require RTT signals favor timing-based
sources as well. Also check for bad TS ECR values to avoid RTT
blow-ups. It has happened on production Web servers.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The first patch consolidates SYNACK and other RTT measurement to use a
central function tcp_ack_update_rtt(). A (small) bonus is now SYNACK
RTT measurement happens after PAWS check, potentially reducing the
impact of RTO seeding on bad TCP timestamps values.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In previous discussions, I tried to find some reasonable heuristics
for delayed ACK, however this seems not possible, according to Eric:
"ACKS might also be delayed because of bidirectional
traffic, and is more controlled by the application
response time. TCP stack can not easily estimate it."
"ACK can be incredibly useful to recover from losses in
a short time.
The vast majority of TCP sessions are small lived, and we
send one ACK per received segment anyway at beginning or
retransmits to let the sender smoothly increase its cwnd,
so an auto-tuning facility wont help them that much."
and according to David:
"ACKs are the only information we have to detect loss.
And, for the same reasons that TCP VEGAS is fundamentally
broken, we cannot measure the pipe or some other
receiver-side-visible piece of information to determine
when it's "safe" to stretch ACK.
And even if it's "safe", we should not do it so that losses are
accurately detected and we don't spuriously retransmit.
The only way to know when the bandwidth increases is to
"test" it, by sending more and more packets until drops happen.
That's why all successful congestion control algorithms must
operate on explicited tested pieces of information.
Similarly, it's not really possible to universally know if
it's safe to stretch ACK or not."
It still makes sense to enable or disable quick ack mode like
what TCP_QUICK_ACK does.
Similar to TCP_QUICK_ACK option, but for people who can't
modify the source code and still wants to control
TCP delayed ACK behavior. As David suggested, this should belong
to per-path scope, since different pathes may want different
behaviors.
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Rick Jones <rick.jones2@hp.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Thomas Graf <tgraf@suug.ch>
CC: David Laight <David.Laight@ACULAB.COM>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Linux sends new unset data during disorder and recovery state if all
(suspected) lost packets have been retransmitted ( RFC5681, section
3.2 step 1 & 2, RFC3517 section 4, NexSeg() Rule 2). One requirement
is to keep the receive window about twice the estimated sender's
congestion window (tcp_rcv_space_adjust()), assuming the fast
retransmits repair the losses in the next round trip.
But currently it's not the case on the first round trip in either
normal or Fast Open connection, beucase the initial receive window
is identical to (expected) sender's initial congestion window. The
fix is to double it.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If the receiver supports DSACK, sender can detect false recoveries and
revert cwnd reductions triggered by either severe network reordering or
concurrent reordering and loss event.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Upon detecting spurious fast retransmit via timestamps during recovery,
use PRR to clock out new data packet instead of retransmission. Once
all retransmission are proven spurious, the sender then reverts the
cwnd reduction and congestion state to open or disorder.
The current code does the opposite: it undoes cwnd as soon as any
retransmission is spurious and continues to retransmit until all
data are acked. This nullifies the point to undo the cwnd because
the sender is still retransmistting spuriously. This patch fixes
it. The undo_ssthresh argument of tcp_undo_cwnd_reductiuon() is no
longer needed and is removed.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactor and relocate various functions or variables to prepare the
undo fix. Remove some unused function arguments. Rename tcp_undo_cwr
to tcp_undo_cwnd_reduction to be consistent with the rest of
CWR related function names.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch series fixes an undo bug in fast recovery: the sender
mistakenly undos the cwnd too early but continues fast retransmits
until all pending data are acked. This also multiplies the SNMP
stat PARTIALUNDO events by the degree of the network reordering.
The first patch prepares the fix by consolidating the accounting
of newly_acked_sacked in tcp_cwnd_reduction(), instead of updating
newly_acked_sacked everytime sacked_out is adjusted. Also pass
acked and prior_unsacked as const type because they are readonly
in the rest of recovery processing.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
case TCP_FIN_WAIT1 can also be simplified by reversing tests
and adding breaks;
Add braces after case and move automatic definitions.
Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
case TCP_SYN_RECV: can have another indentation level removed
by converting
if (acceptable) {
...;
} else {
return 1;
}
to
if (!acceptable)
return 1;
...;
Reflow code and comments to fit 80 columns.
Another pure cleanup patch.
Signed-off-by: Joe Perches <joe@perches.com>
Improved-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove one level of indentation 'introduced' in commit
c3ae62af8e (tcp: should drop incoming frames without ACK flag set)
if (true) {
...
}
@acceptable variable is a boolean.
This patch is a pure cleanup.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Merge net into net-next because some upcoming net-next changes
build on top of bug fixes that went into net.
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is a fix for a bug triggering newly_acked_sacked < 0
in tcp_ack(.).
The bug is triggered by sacked_out decreasing relative to prior_sacked,
but packets_out remaining the same as pior_packets. This is because the
snapshot of prior_packets is taken after tcp_sacktag_write_queue() while
prior_sacked is captured before tcp_sacktag_write_queue(). The problem
is: tcp_sacktag_write_queue (tcp_match_skb_to_sack() -> tcp_fragment)
adjusts the pcount for packets_out and sacked_out (MSS change or other
reason). As a result, this delta in pcount is reflected in
(prior_sacked - sacked_out) but not in (prior_packets - packets_out).
This patch does the following:
1) initializes prior_packets at the start of tcp_ack() so as to
capture the delta in packets_out created by tcp_fragment.
2) introduces a new "previous_packets_out" variable that snapshots
packets_out right before tcp_clean_rtx_queue, so pkts_acked can be
correctly computed as before.
3) Computes pkts_acked using previous_packets_out, and computes
newly_acked_sacked using prior_packets.
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_timeout_skb() was intended to trigger fast recovery on timeout,
unfortunately in reality it often causes spurious retransmission
storms during fast recovery. The particular sign is a fast retransmit
over the highest sacked sequence (SND.FACK).
Currently the RTO timer re-arming (as in RFC6298) offers a nice cushion
to avoid spurious timeout: when SND.UNA advances the sender re-arms
RTO and extends the timeout by icsk_rto. The sender does not offset
the time elapsed since the packet at SND.UNA was sent.
But if the next (DUP)ACK arrives later than ~RTTVAR and triggers
tcp_fastretrans_alert(), then tcp_timeout_skb() will mark any packet
sent before the icsk_rto interval lost, including one that's above the
highest sacked sequence. Most likely a large part of scorebard will be
marked.
If most packets are not lost then the subsequent DUPACKs with new SACK
blocks will cause the sender to continue to retransmit packets beyond
SND.FACK spuriously. Even if only one packet is lost the sender may
falsely retransmit almost the entire window.
The situation becomes common in the world of bufferbloat: the RTT
continues to grow as the queue builds up but RTTVAR remains small and
close to the minimum 200ms. If a data packet is lost and the DUPACK
triggered by the next data packet is slightly delayed, then a spurious
retransmission storm forms.
As the original comment on tcp_timeout_skb() suggests: the usefulness
of this feature is questionable. It also wastes cycles walking the
sack scoreboard and is actually harmful because of false recovery.
It's time to remove this.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_fixup_rcvbuf() contains a loop to estimate initial socket
rcv space needed for a given mss. With large MTU (like 64K on lo),
we can loop ~500 times and consume a lot of cpu cycles.
perf top of 200 concurrent netperf -t TCP_CRR
5.62% netperf [kernel.kallsyms] [k] tcp_init_buffer_space
1.71% netperf [kernel.kallsyms] [k] _raw_spin_lock
1.55% netperf [kernel.kallsyms] [k] kmem_cache_free
1.51% netperf [kernel.kallsyms] [k] tcp_transmit_skb
1.50% netperf [kernel.kallsyms] [k] tcp_ack
Lets use a 100% factor, and remove the loop.
100% is needed anyway for tcp_adv_win_scale=1
default value, and is also the maximum factor.
Refs: commit b49960a05e
("tcp: change tcp_adv_win_scale and tcp_rmem[2]")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add MIB counters for checksum errors in IP layer,
and TCP/UDP/ICMP layers, to help diagnose problems.
$ nstat -a | grep Csum
IcmpInCsumErrors 72 0.0
TcpInCsumErrors 382 0.0
UdpInCsumErrors 463221 0.0
Icmp6InCsumErrors 75 0.0
Udp6InCsumErrors 173442 0.0
IpExtInCsumErrors 10884 0.0
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller