for the full number of tins.
Signed-off-by: Toke Høiland-Jørgensen <t...@toke.dk>
---
This patch is against the for_upstream_4.16 branch. It compiles, but is
otherwise completely untested. No idea if it will fix the build bot
error from the upstream submission, but couldn't think of any other
o
This popped up in my Google Scholar notifications:
https://atlas.informatik.uni-tuebingen.de/~menth/papers/Menth18b.pdf
Basically, they are proposing to permit a queue to accumulate a larger
deficit while empty to allow light users to achieve the same throughput
as heavy users (users being an
Pete Heist writes:
> Whether or not it helps with the upcoming talk or future plans, I’m
> thinking about the most recent experience with trying to upstream Cake
> and how to break the stalemate.
Not sure there is a stalemate, actually. I just ran out of time to do
further revisions; planning
Jonathan Morton writes:
>> On 19 Jun, 2018, at 12:55 pm, Toke Høiland-Jørgensen wrote:
>>
>> Not sure there is a stalemate, actually. I just ran out of time to do
>> further revisions; planning to take that up again. Don't see any reason
>> why we shouldn't be a
Jonathan Morton writes:
> Yes, eBPF does seem to be a good fit for that.
>
> So in summary, the logical flow of a packet should be:
>
> 1: Map dst or src IP to subscriber (eBPF).
> 2: Map subscriber to speed/overhead tier (eBPF).
> 3: (optional) Classify Diffserv (???).
> 4: Enqueue per flow,
Dave Taht writes:
> On Wed, Aug 1, 2018 at 10:25 AM Dave Taht wrote:
>>
>> I wonder if ebpf has opcode space for an invsqrt?
>
> bpf_ktime_get_ns() exists...
>
> one thing that I don't know if bpf can do is read/write the
> skb->tstamp field. The plan would be to rigorously write it (if not
>
Dave Taht writes:
> On Thu, Aug 2, 2018 at 1:04 PM Toke Høiland-Jørgensen wrote:
>>
>> Dave Taht writes:
>>
>> > On Wed, Aug 1, 2018 at 10:25 AM Dave Taht wrote:
>> >>
>> >> I wonder if ebpf has opcode space for an invsqrt?
>> &g
Sebastian Moeller writes:
> Hi Pete
>
>> On Jul 30, 2018, at 11:14, Pete Heist wrote:
>>
>>
>>> On Jul 29, 2018, at 9:14 PM, Toke Høiland-Jørgensen wrote:
>>>>
>>>> Caveats that I know of:
>>>> - Limited to 1024 members
Pete Heist writes:
>> On Jul 30, 2018, at 12:55 PM, Toke Høiland-Jørgensen wrote:
>>>
>>> I believe all you needto do is change the following in scg_cake.c:
>>> #define CAKE_QUEUES (1024)
>>>
>>> Now I heard reports that above a certain num
pat/linux/bpf.h#L
>says you can get at the priority field.
>
>On Sat, Jul 28, 2018 at 10:52 AM Dave Taht wrote:
>>
>> On Sat, Jul 28, 2018 at 10:38 AM Pete Heist wrote:
>> >
>> >
>> > On Jul 28, 2018, at 10:56 AM, Toke Høiland-Jørgensen
>wrot
On 28 July 2018 19:53:58 CEST, Jonathan Morton wrote:
>>> Note that with the existing tc classifier stuff we already added to
>>> Cake, we basically have this already (eBPF can map traffic to tin
>and
>>> flow however it pleases).
>>
>> Sorry, this just jostled in my brain now that I may be
Pete Heist writes:
>> On Jul 30, 2018, at 1:28 PM, Toke Høiland-Jørgensen wrote:
>>
>> Pete Heist writes:
>>
>>> Couldn’t it still be made so now? Not sure of the performance impact
>>> though.
>>
>> It could, but it would tak
Pete Heist writes:
>> On Jul 28, 2018, at 8:12 PM, Toke Høiland-Jørgensen wrote:
>>
>> Priority field sets tin, class sets flow. Both need the qdisc is as its
>> major number, iirc. And both can be set from the same bpf filter which can
>> be run in direct
Dan Siemon writes:
> On Thu, 2018-07-26 at 19:42 +0200, Toke Høiland-Jørgensen wrote:
>> Dan Siemon writes:
>>
>> > I haven't had time to try Cake in this context yet but hope to get
>> > to
>> > that in the next couple months. I believe this wi
Dan Siemon writes:
> Tiny bit of self promotion here but Preseem (https://www.preseem.com)
> is a transparent bridge that leverages HTB/FQ-CoDel to make subscriber
> plan enforcement provide much better QoE. Leaving enforcement up to
> the deep queues in most network equipment has comparably
Dave Taht writes:
> some word from avind's group.
>
> "Thanks Dave for the pointer! Sounds very interesting, so we will
> definitely take a look.
>
> We do have p4 code for AFQ and happy to send it your way. There are
> two crucial components to the AFQ implementation -- one is the switch
>
Dave Taht writes:
> On Wed, Jul 25, 2018 at 3:07 AM Sebastian Moeller wrote:
>>
>> Hi Toke,
>>
>>
>> > On Jul 25, 2018, at 12:02, Toke Høiland-Jørgensen wrote:
>> >
>> > Dave Taht writes:
>> >
>> >> I really want
Dave Taht writes:
> I wanted to look at a few things - cpu usage, 4 different tcps,
> different server schedulers, ecn vs non-ecn, sqm fq_codel simplest.qos
> vs cake, etc, etc. I just did tcp_ndown tests of 128 flows to see what
> happened for starters. I also tried to capture tcp_cwnd, etc,
Dave Taht writes:
> It turns out it's just two ethernets with one, connected to a 2 port
> switch. Not what I wanted. I'd wanted something different from the
> apu2 or edgerouter X to play with, and I know the mvneta driver was
> bql'd.
I bought one of these to play with:
Jonathan Morton writes:
>> On 13 Aug, 2018, at 12:34 am, Jonathan Morton wrote:
>>
>>> On 12 Aug, 2018, at 10:42 pm, Toke Høiland-Jørgensen wrote:
>>>
>>> Yes it does; setting tc_classid is one way to set the flow class that is
>>> read by
>> The next simplest fix is to ignore the flow ID override unless we're
>> in "flows" mode. We can then make valid assumptions about what should
>> go into the host tables.
>>
>> The *right* fix, if we want to maximise functionality, would be to
>> pass the result struct by reference into
Pete Heist writes:
> The eBPF verifier seems fragile to me, where I’d be moving lines of
> code around and getting different error messages in an alien tongue.
Well, it operates on the byte code and errs on the side of safety. I.e.,
if it can't prove your program is safe it is going to reject
On 21 August 2018 23:06:11 CEST, Pete Heist wrote:
>
>> On Aug 21, 2018, at 1:25 PM, Toke Høiland-Jørgensen
>wrote:
>>
>>>> The next simplest fix is to ignore the flow ID override unless
>we're
>>>> in "flows" mode. We can then make val
Pete Heist writes:
>> On Aug 22, 2018, at 8:17 AM, Jonathan Morton wrote:
>>
>>> On 22 Aug, 2018, at 12:06 am, Pete Heist wrote:
>>>
>>> when fq_codel is the qdisc, the eBPF action is only called "once in a while”
>>
>> One difference between fq_codel and Cake is that the former - which
>>
Pete Heist writes:
>> On Aug 21, 2018, at 11:17 PM, Toke Høiland-Jørgensen wrote:
>>>
>>> Well that’s good timing for me as I’m wrapping up a small utility/eBPF
>>> to classify an arbitrary username to either MAC or IP. Here’s the work
>>> in progre
for
the portions that are not set by the filter.
Signed-off-by: Toke Høiland-Jørgensen
---
net/sched/sch_cake.c | 23 +++
1 file changed, 19 insertions(+), 4 deletions(-)
diff --git a/net/sched/sch_cake.c b/net/sched/sch_cake.c
index 35fc7252187c..6bdf6ba06775 100644
--- a/net
Jonathan Morton writes:
>> On 22 Aug, 2018, at 12:51 pm, Pete Heist wrote:
>>
>> "math between pkt pointer and 4294901760 is not allowed"
>
> As a possible clue here, 4294901760 == (2^32) - (2^16).
>
> I suspect both errors are being caused by the call to memcpy(). This
> potentially inlines a
Toke Høiland-Jørgensen writes:
>>Lastly, if anyone has time to review even just a little code for what
>>is or is not good or idiomatic C, post an issue and I’d appreciate it.
>>Yes, I yield to the ‘goto’ proponents when it comes to error handling
>>and resource de-allo
Since CAKE now has three different settings that can be overridden by tc
filters (priority and host and flow hashes), documenting how they work is
probably a good idea.
Signed-off-by: Toke Høiland-Jørgensen
---
man/man8/tc-cake.8 | 55 ++
1 file
Now that CAKE has been accepted upstream, I figured it was a good time
to backport the 'tc class' support. So I did, back to kernel v4.9.
This is in the master branch; anyone feel like testing? With this, the
version of CAKE in the master branch should be identical to the version
that will be in
Kevin Darbyshire-Bryant writes:
>> On 14 Jul 2018, at 22:50, Toke Høiland-Jørgensen wrote:
>>
>> Now that CAKE has been accepted upstream, I figured it was a good time
>> to backport the 'tc class' support. So I did, back to kernel v4.9.
>>
>> This is in
Yeah, I agree that at 1 Gbit you don't need multiple receive queues to
get to line rate. In my 100Gbit tests, I got to 50 Gbps with CAKE (I
should really post some graphs of that), so at really high speeds we
would benefit from being able to run simultaneously on multiple CPUs.
But let's just say
.
sch_cake replaces a combination of iptables, tc filter, htb and fq_codel
in the sqm-scripts, with sane defaults and vastly simpler configuration.
Cake's principal author is Jonathan Morton, with contributions from
Kevin Darbyshire-Bryant, Toke Høiland-Jørgensen, Sebastian Moeller,
Ryan Mounce, Tony A
Kevin Darbyshire-Bryant writes:
>> On 15 Jul 2018, at 10:48, Toke Høiland-Jørgensen wrote:
>>
>> Kevin Darbyshire-Bryant writes:
>>
>>>> On 14 Jul 2018, at 22:50, Toke Høiland-Jørgensen wrote:
>>>>
>>>> Now that CAKE has been acce
Dave Taht writes:
> excluding some painful implementation details
Heh...
-Toke
___
Cake mailing list
Cake@lists.bufferbloat.net
https://lists.bufferbloat.net/listinfo/cake
Dave Taht writes:
> I am curious if the new tc class support would allow for recognizing
> sebastian's desire to have the llc (?) dsl control packets moved to
> the higher priority tin?
>
> Similarly, there were other things in early-sqm (like deprioritizing
> ping, prioritizing dns) that I
Jonathan Morton writes:
>> On 16 Jul, 2018, at 1:41 am, Kevin Darbyshire-Bryant
>> wrote:
>>
>> AFAICT the tin order makes no difference whatsoever these days,
>> indeed the dequeue mechanism picks up from the last tin and spins
>> around rather than starting at either 0 or highest tin each
Jonathan Morton writes:
> On 12:30pm, Mon, 16 Jul 2018 Toke Høiland-Jørgensen, wrote:
>
>> I'm leaning towards making tc filter classification obey tin_order;
>> opinions? :)
>
> Sounds fine to me.
Right, pushed an update to this effect. Will submit it upstream
Stephen Hemminger writes:
> On Mon, 16 Jul 2018 18:39:26 +0200
> Toke Høiland-Jørgensen wrote:
>
>> +#define PRINT_TSTAT(name, attr, fmts, val) do {\
>> +if (GET_TSTAT(0, attr)) { \
>> +
Dave Taht writes:
> On Sat, Jul 21, 2018 at 10:28 AM Georgios Amanakis
> wrote:
>>
>> The previous one was with:
>> net.ipv4.tcp_congestion_control=cubic
>>
>> I retried with:
>> net.ipv4.tcp_congestion_control=reno
>>
>> Georgios
>
> In the fast test this has no effect on the remote server's
A few comments below; will fix the rest.
>> +print_uint(PRINT_JSON, "bandwidth", NULL, bandwidth);
>> +print_string(PRINT_FP, NULL, "bandwidth %s ",
>> sprint_rate(bandwidth, b1));
>> +} else
>> +print_string(PRINT_ANY,
.
sch_cake replaces a combination of iptables, tc filter, htb and fq_codel
in the sqm-scripts, with sane defaults and vastly simpler configuration.
Cake's principal author is Jonathan Morton, with contributions from
Kevin Darbyshire-Bryant, Toke Høiland-Jørgensen, Sebastian Moeller,
Ryan Mounce, Tony A
David Ahern writes:
> On 7/19/18 7:56 AM, Toke Høiland-Jørgensen wrote:
>> sch_cake is intended to squeeze the most bandwidth and latency out of even
>> the slowest ISP links and routers, while presenting an API simple enough
>> that even an ISP can configure it.
>>
&
.
sch_cake replaces a combination of iptables, tc filter, htb and fq_codel
in the sqm-scripts, with sane defaults and vastly simpler configuration.
Cake's principal author is Jonathan Morton, with contributions from
Kevin Darbyshire-Bryant, Toke Høiland-Jørgensen, Sebastian Moeller,
Ryan Mounce, Tony A
Kevin Darbyshire-Bryant writes:
> Hiya Chaps,
>
> Bet that subject woke you up!
>
> This is one for the back burner but I’d like to do it at some point,
> mainly ‘cos my name is in the maintainer field :-)
>
> Ideally I’d like to introduce cake as a kernel patch backport to
> openwrt instead of
David Ahern writes:
> On 7/19/18 4:53 AM, Toke Høiland-Jørgensen wrote:
>> A few comments below; will fix the rest.
>>
>>>> + print_uint(PRINT_JSON, "bandwidth", NULL, bandwidth);
>>>> + print_string(PRINT_FP, N
This is consistent with the other multi-word parameters. Also change the
JSON output to be consistent with way it is formatted for the other
options.
Signed-off-by: Toke Høiland-Jørgensen
---
man/man8/tc-cake.8 | 4 ++--
tc/q_cake.c| 8
2 files changed, 6 insertions(+), 6
Georgios Amanakis writes:
> On Mon, 2018-07-16 at 12:18 +0200, Toke Høiland-Jørgensen wrote:
>> Right, pushed an update to this effect. Will submit it upstream as
>> well
>> once someone has verified that it does what it's supposed to :)
>
> I just verified it work
how they are displayed.
Fix this by adding the missing mapping when reading skb->priority.
Fixes: 83f8fd69af4f ("sch_cake: Add DiffServ handling")
Signed-off-by: Toke Høiland-Jørgensen
---
net/sched/sch_cake.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/sched/
Kevin Darbyshire-Bryant writes:
> Hi All,
>
> I’ve been gently dabbling with wikipedia and as I was playing with ECN
> on a NAS box decided to look at
> https://en.wikipedia.org/wiki/Explicit_Congestion_Notification
>
> The CAKE bit now has a citation and slightly more flashing lights
> around
Dave Taht writes:
> I really wanted cake to always optimize for low latency. I wanted it
> to "just work" at line rate on dsl, on 100mbit, even 10mbit ethernet,
> to work against pause frames, etc, without configuration. I wanted to
> defeat drivers like the mvneta that can do 64k of software
Pete Heist writes:
> I happen to also be working on a bridge setup, but it’s different. For
> one, I used fq_codel on a transparent bridge for a couple years in
> production and it worked well, so I trust it also would for cake.
>
> But now, my neighbor will access the Internet through my CPE
Pete Heist writes:
>> On Sep 6, 2018, at 7:22 PM, Dave Taht wrote:
>>
>> There was a very good paper or two (I think luca co-authored one) that
>> showed that "active flows" were generally measured in the mid 200s in
>> nearly any scenario. I agreed with that which was in part why I felt
>> we
Hi everyone
While investigating a bug report on CAKE[0], I've run into the following
behaviour:
When running CAKE as an ingress shaper on an IFB interface, if the GSO
splitting feature is turned on, TCP throughput will drop dramatically on
6in4 (sit) tunnels running over the interface in
Georgios Amanakis writes:
> Dear All,
>
> I was giving a transparent firewall a try, and wondered whether cake
> can be applied on the interfaces of a bridge. I want to put an extra
> router in-line between clients and the ISP-modem-router. It will have
> two interfaces (eth0 facing wan, eth1
Pete Heist writes:
>> On Jul 6, 2018, at 1:33 PM, Toke Høiland-Jørgensen wrote:
>>
>> AHA! Found the culprit!
>>
>> The bulk dequeue mechanism in sch_generic.c will dequeue a bunch of
>> packets at once, then check if they belong on the same hardware txq
George Amanakis writes:
> I guess the aborts I reported in loop 'k' are normal, just denoting that
> a flow's deficit is <=0.
> If I increase the counter to >10k, I don't see any warnings anymore.
Yeah, some looping is fine, it's the infinite loops we want to avoid ;)
-Toke
Dave Taht writes:
> The word "milestone" doesn't seem adequate. Parsecstone?
>
> Words cannot express how I feel about all your efforts on the
> internet's behalf... But... Thx.
Yes, many thanks to all involved! :D
-Toke
___
Cake mailing list
Pete Heist writes:
> I don’t know if we want to call this an issue, but...
>
> I’m seeing a lockup with cake (and also sfq, but not either pfifo or
> fq_codel), when run over veth devices. Two network namespaces are
> created, one for client and one for server, each with one veth device.
> Netem
ew tweaks to the behaviour of cake based on testing carried out
while writing the paper.
---
Toke Høiland-Jørgensen (8):
sched: Add Common Applications Kept Enhanced (cake) qdisc
sch_cake: Add ingress mode
sch_cake: Add optional ACK filter
netfilter: Add nf_ct_get_tuple_skb glob
ode in sch_cake.
Cc: netfilter-de...@vger.kernel.org
Signed-off-by: Toke Høiland-Jørgensen
---
include/linux/netfilter.h | 11 +++
net/netfilter/core.c | 15 +++
net/netfilter/nf_conntrack_core.c | 36
3 files changed,
on
bandwidth. For this reason, we split GSO segments into their individual
packets iff the shaper is active and configured to a bandwidth <= 1 Gbps.
Signed-off-by: Toke Høiland-Jørgensen
---
net/sched/sch_cake.c | 99 +-
1 file changed, 73 inserti
number of configured priority
tiers.
Signed-off-by: Toke Høiland-Jørgensen
---
net/sched/sch_cake.c | 439 --
1 file changed, 423 insertions(+), 16 deletions(-)
diff --git a/net/sched/sch_cake.c b/net/sched/sch_cake.c
index 633ca1578114..43eeca81b247 100
s and vastly simpler configuration.
CAKE's principal author is Jonathan Morton, with contributions from
Kevin Darbyshire-Bryant, Toke Høiland-Jørgensen, Sebastian Moeller,
Ryan Mounce, Tony Ambardar, Dean Scarff, Nils Andreas Svee, Dave Täht,
and Loganaden Velvindron.
Testing from Pete Heis
bandwidths the effect is negligible at best.
Cc: Yuchung Cheng
Cc: Neal Cardwell
Signed-off-by: Toke Høiland-Jørgensen
---
net/sched/sch_cake.c | 454 ++
1 file changed, 452 insertions(+), 2 deletions(-)
diff --git a/net/sched/sch_cake.c b/net/sched
at higher bandwidths. For this
reason, the feature is turned off by default.
Cc: netfilter-de...@vger.kernel.org
Signed-off-by: Toke Høiland-Jørgensen
---
net/sched/sch_cake.c | 51 --
1 file changed, 49 insertions(+), 2 deletions(-)
diff --git a/net
reported by
the kernel is used.
Signed-off-by: Toke Høiland-Jørgensen
---
net/sched/sch_cake.c | 124 ++
1 file changed, 123 insertions(+), 1 deletion(-)
diff --git a/net/sched/sch_cake.c b/net/sched/sch_cake.c
index 43eeca81b247..199670e1eb94 100644
Jonathan Morton writes:
>> On 6 Jul, 2018, at 4:21 pm, Toke Høiland-Jørgensen wrote:
>>
>>> I'm handling it without using any extra permanent state, just making
>>> use of what's already there. Take a look at latest commit.
>>
>> Yup, that al
Dave Taht writes:
> I am curious as to the cpu and thruput difference between pfifo, htb +
> fqcodel v cake at these insane speeds.
Sure, actual test results at ludicrous speed is on my list :)
-Toke
___
Cake mailing list
Cake@lists.bufferbloat.net
David Miller writes:
> From: Toke Høiland-Jørgensen
> Date: Fri, 06 Jul 2018 17:37:19 +0200
>
>> This patch series adds the CAKE qdisc, and has been split up to ease
>> review.
>>
>> I have attempted to split out each configurable feature into its own patch.
Pv4 and IPv6
gso_segment handlers, after they modify the network_header.
Signed-off-by: Toke Høiland-Jørgensen
---
net/ipv4/af_inet.c |1 +
net/ipv6/ip6_offload.c |1 +
2 files changed, 2 insertions(+)
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 20fda8fb8ffd..1fbe2f815474 100
Stephen Hemminger <step...@networkplumber.org> writes:
> On Sun, 11 Feb 2018 18:26:18 +0100
> Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>
>> This updates tc to understand the updated cake xstats structure (which
>> splits out the tin stats in a separ
Toke Høiland-Jørgensen <t...@toke.dk> writes:
> Stephen Hemminger <step...@networkplumber.org> writes:
>
>> On Sun, 11 Feb 2018 18:26:18 +0100
>> Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>
>>> This updates tc to understand the updated cake x
Kevin Darbyshire-Bryant writes:
> Archer c7 v2. master branch of openwrt
Ah, great; I actually have one of those sitting on my desk that I could
potentially reflash without breaking anything too important.
In the meantime; do you get the same weird output on the
Kevin Darbyshire-Bryant <ke...@darbyshire-bryant.me.uk> writes:
>> On 8 Mar 2018, at 11:09, Kevin Darbyshire-Bryant
>> <ke...@darbyshire-bryant.me.uk> wrote:
>>
>>
>>
>>> On 8 Mar 2018, at 10:57, Toke Høiland-Jørgensen <t...@toke.dk>
Thank you for the patch! :)
> Signed-off-by: Kevin Darbyshire-Bryant
> ---
> tc/q_cake.c | 15 ---
> 1 file changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/tc/q_cake.c b/tc/q_cake.c
> index 44cadb63..f888bd2a 100644
> --- a/tc/q_cake.c
> +++
Kevin Darbyshire-Bryant <ke...@darbyshire-bryant.me.uk> writes:
>> On 11 Mar 2018, at 20:50, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>
>> Kevin Darbyshire-Bryant <l...@darbyshire-bryant.me.uk> writes:
>>
>>> Signed-off-by: Ke
Stephen Hemminger writes:
>> > Using the ‘PRId64’ macro won’t work because print_int is using ‘int’
>> > type internally whereas print_uint uses ‘uint64_t’ internally. So the
>> > format string has to have knowledge of the internal format, *but*
>> > there’s no clue
Dave Taht writes:
> And while I'm breaking things, the link_ms field is entirely unused in
>
> struct tc_cake_traffic_stats {
> __u32 packets;
> __u32 link_ms;
> __u64 bytes;
> };
what was that field supposed to contain?
-Toke
Dave Taht writes:
> While it seems unlikely we'll get to shaping past 40Gbit any time
> soon, perhaps changing the API to 64 bits to specify bandwidth will
> avoid problems in 2024, or with (one day) offloaded hardware.
Might as well. There is not really much of a
Jonathan Morton <chromati...@gmail.com> writes:
>> On 6 Mar, 2018, at 11:06 pm, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>
>> ...on the iproute2 side the only
>> thing missing before we can attempt an upstream submission is an update
>> of the man
Toke Høiland-Jørgensen <t...@toke.dk> writes:
> Kevin Darbyshire-Bryant <ke...@darbyshire-bryant.me.uk> writes:
>
>> Hi All,
>>
>>
>>
>>> On 7 Mar 2018, at 08:50, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>>
>>> Jonat
Kevin Darbyshire-Bryant <ke...@darbyshire-bryant.me.uk> writes:
> Hi All,
>
>
>
>> On 7 Mar 2018, at 08:50, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>
>> Jonathan Morton <chromati...@gmail.com> writes:
>>
>>>> On 6 M
Jonathan Morton writes:
>> On 1 Mar, 2018, at 1:06 pm, Sebastian Moeller wrote:
>>
>> BTW, my testing so far with the latest tc-adv did not result in any crashes,
>> but I also did not research whether the overhead account behaves like
>> expected...
>
Sebastian Moeller writes:
>> diff --git a/tc/q_cake.c b/tc/q_cake.c
>> index e21552e8..95301b41 100644
>> --- a/tc/q_cake.c
>> +++ b/tc/q_cake.c
>> @@ -243,12 +243,22 @@ static int cake_parse_opt(struct qdisc_util *qu, int
>> argc, char **argv,
>>/* Typical
Kevin Darbyshire-Bryant writes:
> There were some useful stats column re-alignment changes as well,
> wonder if you got those?
Probably not, as that code is not longer directly diff'able,
unfortunately... The json-related changes were fairly intrusive...
Kevin Darbyshire-Bryant <ke...@darbyshire-bryant.me.uk> writes:
>> On 7 Mar 2018, at 10:31, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>
>>>
>>
>> Please don't put something different into LEDE than what we're working
>> on upstreaming. It
Kevin Darbyshire-Bryant writes:
> I don’t the column alignment can be correct because the print lines don’t
> include a leading space, so columns can run into each other.
>
> fprintf(f, "%12u", tst->unresponse_flows);
> v
> fprintf(f, " %12u",
Dave Taht writes:
> So it sounds like all the individual pieces are sync'd again?
>
> 4.16-rc4 is out, so there are several weeks left to make the net-next
> window. I note that I don't feel strongly (never had) that I should be
> the one to make the net-next submission,
Pete Heist writes:
> For what it’s worth, that’s what I also saw testing Cake on the APU2
> late last year, and the ER-X platform earlier. I actually never knew
> that Cake used less CPU at some point. Sorry for no supporting
> detail... :)
Anecdotal supporting evidence is
Jonathan Morton <chromati...@gmail.com> writes:
>> On 11 Apr, 2018, at 6:24 pm, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>
>> So, um, did we cram so many features into Cake that it no longer uses
>> less CPU? Can anyone confirm these results?
>
Y via Cake writes:
> From: Y
> Subject: Re: [Cake] A few puzzling Cake results
> To: cake@lists.bufferbloat.net
> Date: Tue, 17 Apr 2018 21:05:12 +0900
>
> Hi.
>
> Any certain fomula of fq_codel flow number?
Well, given N active bulk flows
Luca Muscariello writes:
> I'm not sure that the quantum correction factor is correct.
No, you're right, there's an off-by-one error. It should be:
R_s < R / ((L/L_s)(N+1) + 1)
-Toke
___
Cake mailing list
Jonathan Morton <chromati...@gmail.com> writes:
>> On 17 Apr, 2018, at 12:42 pm, Toke Høiland-Jørgensen <t...@toke.dk> wrote:
>>
>> - The TCP RTT of the 32 flows is *way* higher for Cake. FQ-CoDel
>> controls TCP flow latency to around 65 ms, while for
Jonas Mårtensson <martensson.jo...@gmail.com> writes:
> On Wed, Apr 18, 2018 at 1:25 PM, Toke Høiland-Jørgensen <t...@toke.dk>
> wrote:
>
>> Toke Høiland-Jørgensen <t...@toke.dk> writes:
>>
>> > Jonathan Morton <chromati...@gmail.com> writes:
Toke Høiland-Jørgensen <t...@toke.dk> writes:
> Is anyone actually using the LLT diffserv setting? The draft describing
> it seems to have expired ages ago:
>
> https://datatracker.ietf.org/doc/draft-you-tsvwg-latency-loss-tradeoff/
Also, adding to the diffserv thing, rig
Jonathan Morton writes:
>>> I'm saying that there's a tradeoff between intra-flow induced latency and
>>> packet loss, and I've chosen 4 MTUs as the operating point.
>>
>> Is there a reason for picking 4 MTUs vs 2 MTUs vs 2 packets, etc?
>
> To be more precise, I'm using
Jonathan Morton writes:
>> This is why I think that any fix that tries to solve this problem in
>> the queueing system should be avoided. It does not solve the real
>> problem (overload) and introduces latency.
>
> Most people, myself included, prefer systems that degrade
Pete Heist writes:
>> On Apr 21, 2018, at 6:44 PM, Jonathan Morton wrote:
>>>
>>> Improved intra-flow latency could be useful for HTTP/2 or other things, but
>>> I don’t see how the current llt mode helps most ordinary people with that,
>>> when it
Y via Cake writes:
> From: Y
> Subject: Re: [Cake] Diffserv LLT mode
> To: cake@lists.bufferbloat.net
> Date: Sat, 21 Apr 2018 21:40:48 +0900
>
> Hi.
>
> mine stat doen't inclued more.
> sad.
Did you upgrade the kernel module without
201 - 300 of 578 matches
Mail list logo