n Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/net/tcp.h| 4
net/ipv4/tcp_cong.c | 2 +-
net/ipv4/tcp_input.c | 17 ++---
3 files c
older form of the tcp_skb_cb size check. Doing so allows TCP to use 4
more bytes of the skb->cb[] space.
Fixes: b4772ef879a8 ("net: use common macro for assering skb->cb[] available
size in protocol families")
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yega
end_phase_rtt, ack_phase_rtt)
In large-scale experiments, this filtering approach incorporating
send_phase_rtt is effective at avoiding bandwidth overestimates due to
ACK compression or stretched ACKs.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-o
-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/uapi/linux/pkt_sched.h | 2 ++
net/sched/sch_fq.c | 22 +++---
2 files
use the windowed
min/max filter code.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/linux/tcp.h | 5 ++--
include/net/tcp.h| 2
The TCP CUBIC module already uses 64 bytes.
The upcoming TCP BBR module uses 88 bytes.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/net
Export tcp_mss_to_mtu(), so that congestion control modules can use
this to help calculate a pacing rate.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
increasing or decreasing.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/linux/win_minmax.h | 37 +
lib/Makefile | 2
expand the TCP send buffer. For congestion
control modules that do not specificy this callback, TCP continues to
use the previous default of 2.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
l for applications that
want to know the current throughput the TCP connection is seeing,
e.g. adaptive bitrate video streaming. It can also be very useful for
debugging or troubleshooting.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukk
losses when a policer kicks
in.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/linux/tcp.h | 1 +
net/ipv4/tcp_input.c | 25
max
receive window offered by the TCP protocol is 2^(16+14) = 2^30. So we
only need 30 bits for the tx.in_flight used by NV.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Has
per TSO skb, in case the congestion control
module has a different notion of the best floor for TSO skbs for
the connection right now. For very low-rate paths or policed
connections it can be appropriate to use smaller TSO skbs.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
further
research.
Test results, performance evaluations, feedback, and BBR-related
discussions are very welcome in the public e-mail list for BBR:
https://groups.google.com/forum/#!forum/bbr-dev
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: N
This would create a
compilation error for tcp_cdg.c, which defines its own minmax
struct. To avoid this naming conflict (and potentially others in the
future), this commit renames the version used in tcp_cdg.c to
cdg_minmax.
Signed-off-by: Soheil Hassas Yeganeh
Signed-off-by: Neal Cardwell
Signed-o
tcp_skb_cb size check in tcp_init()
Eric Dumazet (2):
net_sched: sch_fq: add low_rate_threshold parameter
tcp: switch back to proper tcp_skb_cb size check in tcp_init()
Neal Cardwell (8):
lib/win_minmax: windowed min or max estimator
tcp: use windowed min filter library for TCP min_rtt estima
On Fri, Sep 16, 2016 at 5:38 PM, kbuild test robot wrote:
> Hi Yuchung,
>
> [auto build test WARNING on net-next/master]
> All warnings (new ones prefixed by >>):
>
>In file included from net/ipv4/route.c:103:0:
>>> include/net/tcp.h:769:11: warning: 'packed' attribute ignored for field of
>>
On Fri, Sep 16, 2016 at 11:56 PM, kbuild test robot wrote:
>
>>> net/ipv4/tcp.c:2794:3: note: in expansion of macro 'do_div'
> do_div(rate, intv);
> ^~
>In file included from arch/arm/include/asm/div64.h:126:0,
> from include/linux/kernel.h:142,
>
On Fri, Sep 16, 2016 at 3:21 PM, kbuild test robot wrote:
> All errors (new ones prefixed by >>):
>
>>> net/ipv4/tcp_cdg.c:59:8: error: redefinition of 'struct minmax'
> struct minmax {
>^~
>In file included from include/linux/tcp.h:22:0,
> from include/
The TCP CUBIC module already uses 64 bytes.
The upcoming TCP BBR module uses 88 bytes.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/net
per TSO skb, in case the congestion control
module has a different notion of the best floor for TSO skbs for
the connection right now. For very low-rate paths or policed
connections it can be appropriate to use smaller TSO skbs.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
n Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/net/tcp.h| 4
net/ipv4/tcp_cong.c | 2 +-
net/ipv4/tcp_input.c | 17 ++---
3 files c
max
receive window offered by the TCP protocol is 2^(16+14) = 2^30. So we
only need 30 bits for the tx.in_flight used by NV.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Has
end_phase_rtt, ack_phase_rtt)
In large-scale experiments, this filtering approach incorporating
send_phase_rtt is effective at avoiding bandwidth overestimates due to
ACK compression or stretched ACKs.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-o
-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/uapi/linux/pkt_sched.h | 2 ++
net/sched/sch_fq.c | 22 +++---
2 files
further
research.
Test results, performance evaluations, feedback, and BBR-related
discussions are very welcome in the public e-mail list for BBR:
https://groups.google.com/forum/#!forum/bbr-dev
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: N
d-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/net/tcp.h | 2 ++
net/ipv4/tcp_output.c | 15 +--
2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/in
l for applications that
want to know the current throughput the TCP connection is seeing,
e.g. adaptive bitrate video streaming. It can also be very useful for
debugging or troubleshooting.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukk
losses when a policer kicks
in.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/linux/tcp.h | 1 +
net/ipv4/tcp_input.c | 25
Export tcp_mss_to_mtu(), so that congestion control modules can use
this to help calculate a pacing rate.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
expand the TCP send buffer. For congestion
control modules that do not specificy this callback, TCP continues to
use the previous default of 2.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
use the windowed
min/max filter code.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/linux/tcp.h | 5 ++--
include/net/tcp.h| 2
increasing or decreasing.
Signed-off-by: Van Jacobson
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Nandita Dukkipati
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
---
include/linux/win_minmax.h | 37 +
lib/Makefile | 2
Dumazet (1):
net_sched: sch_fq: add low_rate_threshold parameter
Neal Cardwell (8):
lib/win_minmax: windowed min or max estimator
tcp: use windowed min filter library for TCP min_rtt estimation
tcp: count packets marked lost for a TCP connection
tcp: allow congestion control module to r
KB/s for single TCP
> connection and kernel 4.3 and above (including 4.8-rc4) shows bandwidth
> ~100KB/s.
> That is caused by stalled cwnd when cwnd equals ssthresh. This patch
> fixes it by proper increasing cwnd in this case.
>
> Signed-off-by: Artem Germanov
Acked-by: Neal Cardwell
neal
Sorry, there's a typo in the subject line: that should be "net" rather
than "next" (I'm proposing "net" since it's a bug fix).
Looks like "git am" strips this mistake, but I'm happy to resubmit if it helps.
thanks,
neal
e it sends
a SYN/ACK for the SYN/data.
Reported-by: Yuchung Cheng
Signed-off-by: Neal Cardwell
Signed-off-by: Yuchung Cheng
Signed-off-by: Eric Dumazet
Signed-off-by: Soheil Hassas Yeganeh
Fixes: 168a8f58059a ("tcp: TCP Fast Open Server - main code path")
---
net/ipv4/tcp_fastopen.c |
like collapsing
> receive queues or partial ofo pruning.
>
> Signed-off-by: Eric Dumazet
> Cc: Yuchung Cheng
> Cc: Neal Cardwell
> ---
> include/net/tcp.h |1 +
> net/ipv4/tcp_ipv4.c | 33 +
> net/ipv6/tcp_ipv6.c |5 +
increase the parent (listener) sk_drops.
>
> Signed-off-by: Eric Dumazet
> ---
Acked-by: Neal Cardwell
Thanks, Eric!
neal
+0 < . 1:11(10) ack 1 win 156
> // check that window is properly scaled !
> +0 > . 1:1(0) ack 1 win 226
>
>
> Signed-off-by: Eric Dumazet
> Cc: Yuchung Cheng
> Cc: Neal Cardwell
> ---
Acked-by: Neal Cardwell
Nice.
neal
ns only a part of the out of order queue in order
> to meet the memory constraints.
>
> Signed-off-by: Eric Dumazet
> Cc: Neal Cardwell
> Cc: Yuchung Cheng
> Cc: Soheil Hassas Yeganeh
> Cc: C. Stephen Gun
> Cc: Van Jacobson
> ---
Nice.
Acked-by: Neal Cardwell
neal
change in slow path, but why not ? ;)
>
> Signed-off-by: Eric Dumazet
> ---
Acked-by: Neal Cardwell
neal
found by Marco Grassi thanks to syzkaller.
>
> Fixes: 6859d49475d4 ("[TCP]: Abstract tp->highest_sack accessing & point to
> next skb")
> Reported-by: Marco Grassi
> Signed-off-by: Eric Dumazet
> Cc: Ilpo Järvinen
> Cc: Yuchung Cheng
> Cc: Neal Cardwell
Acked-by: Neal Cardwell
Thanks, Eric.
neal
_rcv+0x122/0x220
> [] tcp_v4_rcv+0x9a6/0xa80
> [] ip_local_deliver_finish+0x9a/0x200
> [] ip_local_deliver+0x64/0xd0
> [] ? ip_rcv_finish+0x350/0x350
> [] ip_rcv_finish+0x151/0x350
>
> [...]
>
> This upstream patch seems to address this issue:
>
> commit d88270e
-off-by: Soheil Hassas Yeganeh
> Suggested-by: Neal Cardwell
Acked-by: Neal Cardwell
Thanks, Soheil.
neal
_select_initial_window().
>
> To fix this, consider the maximum of net.ipv4.tcp_rmem[2],
> net.core.rmem_max and socket's initial buffer space.
>
> This part of the code does not have git history and as a
> result this patch does not have a `Fixes:` tag.
>
> Signed-off-by: S
-up to:
> Subject: tcp: make challenge acks less predictable
>
> Cc: Eric Dumazet
> Cc: David S. Miller
> Cc: Neal Cardwell
> Cc: Yuchung Cheng
> Cc: Yue Cao
> Signed-off-by: Jason Baron
> ---
> net/ipv4/tcp_input.c | 39 ++-
> 1 file changed, 22 insertions(+), 17 deletions(-)
Sounds good to me.
Acked-by: Neal Cardwell
thanks,
neal
-by: Eric Dumazet
> Suggested-by: Linus Torvalds
> Cc: Yuchung Cheng
> Cc: Neal Cardwell
Acked-by: Neal Cardwell
Thanks, Eric!
neal
On Tue, Jun 28, 2016 at 5:54 PM, Seymour, Shane M wrote:
>> From: Eric Dumazet [mailto:eric.duma...@gmail.com]
...
>> Anyway, your patch is reversed.
>
> I'm not sure what you mean by reversed, I didn't change the direction
> of the test in the code just what it's being compared against and so
> i
Eric Dumazet
> Reported-by: Arjun V
> Tested-by: Arjun V
> ---
> net/ipv4/tcp_output.c | 7 ++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
Acked-by: Neal Cardwell
Thanks, Eric!
neal
On Mon, Jun 27, 2016 at 9:27 AM, Eric Dumazet wrote:
> Excellent idea ;)
>
> Here is the v2 patch.
>
> diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> index 8bd9911fdd16..e00e972c4e6a 100644
> --- a/net/ipv4/tcp_output.c
> +++ b/net/ipv4/tcp_output.c
> @@ -2751,7 +2751,7 @@ void tcp_x
On Mon, Jun 27, 2016 at 8:46 AM, Eric Dumazet wrote:
> Patch would be :
>
> diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
> index 8bd9911fdd16..3587efe22864 100644
> --- a/net/ipv4/tcp_output.c
> +++ b/net/ipv4/tcp_output.c
> @@ -2784,6 +2784,10 @@ void tcp_xmit_retransmit_queue(struc
uct tcp_dctcp_info, in which case the
TCP_CC_INFO code might accidentally copy uninitialized bytes from the
stack.
Signed-off-by: Neal Cardwell
Signed-off-by: Soheil Hassas Yeganeh
Signed-off-by: Eric Dumazet
---
net/ipv4/tcp_dctcp.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/
der to avoid spoofing attacks.
>
>> Signed-off-by: Pau Espin Pedrol
Acked-by: Neal Cardwell
Tested-by: Neal Cardwell
I verified that it passes the following packetdrill test, which is
based on the script posted by Pau in a recent thread on this topic.
Thanks!
neal
// Test that w
The functionality seems OK to me, though there are some
style/formatting issues, which checkpatch.pl picks up:
> ./scripts/checkpatch.pl
> net-next-tcp-accept-RST-if-SEQ-matches-right-edge-of-right-most-SACK-block.patch
WARNING: line over 80 characters
#73: FILE: net/ipv4/tcp_input.c:5199:
+
On Fri, Jun 3, 2016 at 3:44 PM, Neal Cardwell wrote:
> On Fri, Jun 3, 2016 at 12:24 PM, Eric Dumazet wrote:
>> On Fri, 2016-06-03 at 11:45 -0400, Neal Cardwell wrote:
>>> But I would also vote to tighten up the proposed logic slightly, and
>>> only check the seq of t
On Fri, Jun 3, 2016 at 12:24 PM, Eric Dumazet wrote:
> On Fri, 2016-06-03 at 11:45 -0400, Neal Cardwell wrote:
>> But I would also vote to tighten up the proposed logic slightly, and
>> only check the seq of the incoming RST against the right edge of the
>> *right-most* SAC
On Fri, Jun 3, 2016 at 11:13 AM, Eric Dumazet wrote:
> I have no strict opinion on this.
>
> It seems to me that checking at most 4 right edges (at least in current
> linux implementation) is not adding a huge risk, and allows for better
> interoperability.
>
> I vote for no extra sysctl.
I vote
end back an ACK.
>
> This patch adds a rate limitation, so that the effect of such
> attacks is limited to ingress only.
>
> We roughly double our ability to absorb such attacks.
Thanks, Eric!
Acked-by: Neal Cardwell
neal
On Tue, Feb 23, 2016 at 8:57 AM, wrote:
> I published example pcap file under following link:
>
> https://www.dropbox.com/s/v8375ub16seyt1a/test7.pcap?dl=0
>
> I hope it is possible to download it without creating dropbox account.
Thanks for the trace. It looks like for the first 0.8 seconds of
On Tue, Feb 23, 2016 at 6:19 AM, wrote:
> My question is: what causes such dynamic change in
> the window size (while transferring data)?
> Is it some kernel parameter wrong set or something like this?
That sounds like TCP receive buffer auto-tuning (also called "Dynamic
right-sizing (DRS)" (Fisk
cket arrived.
>
> Before queuing a packet, TCP has to set skb->dev to NULL as the device
> could disappear before packet is removed from the queue.
>
> Fix this issue by using the net pointer provided by the socket (being a
> timewait or a request socket).
Thanks, Eric!
Acked-by: Neal Cardwell
neal
On Thu, Dec 10, 2015 at 1:51 AM, Per Hurtig wrote:
>
>> On 08 Dec 2015, at 14:47, Eric Dumazet wrote:
>>
>> On Tue, 2015-12-08 at 10:19 +0100, Per Hurtig wrote:
>>
>>> +static u32 tcp_unsent_pkts(const struct sock *sk, u32 ulimit)
>>> +{
>>> +struct sk_buff *skb = tcp_send_head(sk);
>>> +
On Fri, Oct 23, 2015 at 4:50 PM, Bendik Rønning Opstad
wrote:
>@@ -2409,6 +2412,15 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
...
> + case TCP_RDB:
> + if (val < 0 || val > 1) {
> + err = -EINVAL;
> + } else {
> +
On Wed, Oct 14, 2015 at 5:28 AM, Andrew Shewmaker wrote:
>> diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
>> index 38743e5..e177386 100644
>> --- a/net/ipv4/tcp_input.c
>> +++ b/net/ipv4/tcp_input.c
> ...
>> @@ -2961,7 +3028,7 @@ void tcp_synack_rtt_meas(struct sock *sk, struct
>> requ
On Sat, Oct 3, 2015 at 6:46 PM, Ben Greear wrote:
>
>
> On 10/03/2015 09:29 AM, Neal Cardwell wrote:
>>
>> On Fri, Oct 2, 2015 at 8:21 PM, Ben Greear
>> wrote:
>>>
>>> Gah, seems 'cubic' related. That is the default tcp cong ctrl
>>
On Fri, Oct 2, 2015 at 8:21 PM, Ben Greear wrote:
> Gah, seems 'cubic' related. That is the default tcp cong ctrl
> I was using (same in 3.17, for that matter).
There have been recent changes to CUBIC that may account for this. If
you could repeat your test with more instrumentation, eg "nstat",
opposite :
>
> Once TSopt has been successfully negotiated, that is both and
>contain TSopt, the TSopt MUST be sent in every non-
> segment for the duration of the connection, and SHOULD be sent in an
>segment (see Section 5.2 for details)
>
> Fixes: 7faee5c0d514 (&q
.1).
>
> Cc: Andreas Petlund
> Cc: Carsten Griwodz
> Cc: Jonas Markussen
> Cc: Kenneth Klette Jonassen
> Cc: Mads Johannessen
> Signed-off-by: Bendik Rønning Opstad
> ---
Acked-by: Neal Cardwell
Tested-by: Neal Cardwell
I ran all the Google packetdrill test cases wit
On Wed, Sep 23, 2015 at 10:46 AM, Bendik Rønning Opstad
wrote:
>> On Tue, Sep 22, 2015 at 2:02 PM, Neal Cardwell wrote:
>> > More generally, my sense is that we should tweak the is_cwnd_limited code
>> > to
>> > shift from saying "set is_cwnd_limit
On Tue, Sep 22, 2015 at 2:02 PM, Neal Cardwell wrote:
> More generally, my sense is that we should tweak the is_cwnd_limited code to
> shift from saying "set is_cwnd_limited to true iff the cwnd is known to be
> limiting transmits" to saying "set is_cwnd_limited to
> I'll describe two example scenarios in detail. In both scenarios we are in
> congestion avoidance after experiencing loss. Nagle is disabled.
Thanks for the detailed follow-up! And thanks, Eric, for the packetdrill
script!
This looks like an issue of how to deal with the case when we run out of
On Fri, Sep 18, 2015 at 7:38 PM, Bendik Rønning Opstad
wrote:
>
> Application limited streams such as thin streams, that transmit small
> amounts of payload in relatively few packets per RTT, are prevented from
> growing the CWND after experiencing loss. This leads to increased
> sojourn times for
On Mon, Sep 14, 2015 at 6:27 AM, Jovi Zhangwei wrote:
>
> Hi Near,
>
> After several days testing on your patch, our system crashed. Dmesg attached.
Jovi -- Sorry about that... thank you for the testing and the data point.
neal
--
To unsubscribe from this list: send the line "unsubscribe netdev"
m the check
> if/when receive window opens.
...
> Signed-off-by: Eric Dumazet
> Cc: Neal Cardwell
> Cc: Yuchung Cheng
> ---
Acked-by: Neal Cardwell
Looks good to me. Thanks, Eric!
neal
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of
ect the conservative ratio
> as soon as cwnd >= ssthresh/2 :
>
> - After cwnd reduction, it is safer to ramp up more slowly,
> as we approach optimal cwnd.
> - Initial ramp up (ssthresh == INFINITY) still allows doubling
> cwnd every other RTT.
>
> Signed-off-by: Eric Duma
On Thu, Aug 20, 2015 at 1:08 PM, Eric Dumazet wrote:
> From: Eric Dumazet
>
> slow start after idle might reduce cwnd, but we perform this
> after first packet was cooked and sent.
>
> With TSO/GSO, it means that we might send a full TSO packet
> even if cwnd should have been reduced to IW10.
>
>
On Mon, Aug 10, 2015 at 2:10 PM, Jovi Zhangwei wrote:
>
> Ping?
>
> We saw a lot of this warnings in our production system. It would be
> great appreciate if someone can give us the fix on this warnings. :)
What is your net.ipv4.tcp_mtu_probing setting? If 1, have you tried
setting it to 0? Previ
his patch removes the limitation, as we did for CWR state
> in commit a0ea700e409 ("tcp: tso: allow CA_CWR state in
> tcp_tso_should_defer()")
>
> Neal Cardwell had a concern about limited transmit so
> Yuchung conducted experiments on GFE and found nothing
> worth ad
On Wed, Jul 22, 2015 at 2:50 AM, Yuchung Cheng wrote:
> On Tue, Jul 21, 2015 at 9:21 PM, Lawrence Brakmo wrote:
>> This is a request for comments.
>>
>> TCP-NV (New Vegas) is a major update to TCP-Vegas. An earlier version of
>> NV was presented at 2010's LPC (slides). It is a delayed based
>> co
n without having to modify existing congestion control
>> modules (tcp_nv in particular needs bytes in flight when packet
>> was sent).
>>
>> This was proposed by Neal Cardwell in his comments to the tcp_nv patch.
>
> Are you sure Neal suggested to pass a struct as argument
On Thu, Jul 9, 2015 at 12:01 PM, Eric Dumazet wrote:
>
> From: Eric Dumazet
>
> After commit 900f65d361d3 ("tcp: move duplicate code from
> tcp_v4_init_sock()/tcp_v6_init_sock()"), we no longer
> need to export tcp_init_xmit_timers()
>
> Signed-off-by: Eric Du
> diff --git a/include/linux/tcp.h b/include/linux/tcp.h
> index 48c3696..05e0da5 100644
> --- a/include/linux/tcp.h
> +++ b/include/linux/tcp.h
> @@ -254,6 +254,10 @@ struct tcp_sock {
> u32 lost_out; /* Lost packets */
> u32 sacked_out; /* SACK'd
On Sat, Jun 6, 2015 at 11:53 AM, Eric Dumazet wrote:
> From: Eric Dumazet
...
> This patch adds a new SOL_IP socket option, asking kernel to ignore
> the 0 port provided by application in bind(IP, port=0) and only
> remember the given IP address.
...
> This new feature is available for both IPv4
On Thu, Jun 4, 2015 at 12:35 PM, Grant Zhang wrote:
> Hi Neal,
>
> Unfortunately with the patch we still see the same stack trace. Attached is
> the TcpExtTCPSACKReneging with the patch, captured with 60 seconds interval.
> Its value is incremented at an similar speed as before, about 30/minute.
>
On Sat, May 30, 2015 at 2:52 PM, Grant Zhang wrote:
> Thank you Neal. Most likely I will test the patch on Monday and report
> back the result.
>
> As for the TcpExtTCPSACKReneging counter, attached is the captured
> counter value on a 1-second interval for 10 minutes.
OK, great. Those TcpExtTCPS
On Fri, May 29, 2015 at 3:53 PM, Grant Zhang wrote:
> Hi Neal,
>
> I will be more happy to test the patch. Please send it my way.
Great. Thank you so much for being willing to do this. Attached is a
patch for testing. I generated it and tested it relative to Linux
v3.14.39, since your stack trace
On Fri, May 29, 2015 at 3:21 PM, Grant Zhang wrote:
> We have multiple machines running into the following trace repeatedly. The
> trace shows up every couple of seconds on our production machines.
>
...
> May 29 18:14:04 cache-fra1230 kernel:[3080455.796188] []
> tcp_fragment+0x2e4/0x2f0
> May
Fixes: 55d8694fa82c ("net: tcp: assign tcp cong_ops when tcp sk is created")
Cc: Florian Westphal
Cc: Daniel Borkmann
Cc: Glenn Judd
Cc: Stephen Hemminger
Signed-off-by: Neal Cardwell
Signed-off-by: Eric Dumazet
Signed-off-by: Yuchung Cheng
---
include/net/inet_connection_s
> We can then revert 843925f33fcc293d80acf2c5c8a78adf3344d49b
> ("tcp: Do not apply TSO segment limit to non-TSO packets")
> and save few cpu cycles in fast path.
Thanks, Eric! This is a nice clean-up/simplification.
Acked-by: Neal Cardwell
neal
--
To unsubscribe from this list: send the line "un
based retransmits)
>
> A followup patch adds a new SNMP counter, as it would have helped a lot
> diagnosing this issue.
>
> Signed-off-by: Eric Dumazet
> Signed-off-by: Yuchung Cheng
Acked-by: Neal Cardwell
neal
--
To unsubscribe from this list: send the line &qu
Reported-by: kbuild test robot
> Signed-off-by: Eric Dumazet
Acked-by: Neal Cardwell
neal
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
y have a socket file descriptor,
> instead of playing complex netlink games.
>
> Sample usage would be :
>
> union tcp_cc_info info;
> socklen_t len = sizeof(info);
>
> if (getsockopt(fd, SOL_TCP, TCP_CC_INFO, &info, &len) == -1)
>
> Signed-off-
info(), so that following patch
> can reuse this common infrastructure.
>
> Signed-off-by: Eric Dumazet
> Cc: Yuchung Cheng
> Cc: Neal Cardwell
> ---
Acked-by: Neal Cardwell
Thanks, Eric!
neal
--
To unsubscribe from this list: send the line "unsubscribe netd
301 - 394 of 394 matches
Mail list logo
