Please keep netdev and myself CCed.
Frithjof Hammer wrote:
Does this patch help?
A further examiniation:
[...]
printk (fri: mein type %x\n,dev-type);
switch (dev-type) {
[...]
shows this:
[EMAIL PROTECTED]:/usr/src/linux-source-2.6.21# dmesg | grep fri
fri: mein
Frithjof Hammer wrote:
Sorry, I didnt follow the thread - what is the goal to be achieved with
the setup?
A simple ingress shaping on ppp0 (PPPOE DSL line). I want to replace my old
imq ingress shaper in favor of ifb. My former script used iptables marks to
classify the packets. My
Frithjof Hammer wrote:
My goal is to setup an ingress traffic shaping on my PPPOE DSL line with ifb.
My old imq stuff used iptables marks (like 'iptables -t mangle -A
PREROUTING -p tcp --sport 22 -m length --length :500 -j MARK --set-mark 31')
to classify the traffic and since i am lazy,
[Please keep me in CC/To, I don't read lartc often]
Mario Antonio Garcia wrote:
I used to get an average of 18900kbit.
My hope was that these new patches would bring better accuracy.
Well, you're up from 94.5% to 99.95%, so they seem to do :)
Notice the 24 cburst. I am just trying to
On Wed, 29 Aug 2007, Mario Antonio Garcia wrote:
I wonder if somebody has got good results (accurate shaping) using 2.6.22?
I am testing with 2.6.22.1, and I haven't been able to get accurate shaping.
For instance, I tried:
$TC qdisc add dev eth0 root handle 1: htb default 1
$TC
Daniel Harold L. wrote:
Dear all,
First, sorry for my bad English ..
To night one of my client is the victim of UDP attack from internet. It's
tons
of UDP packets from internet with destination to port 80. But when I look at
class of that victim client, the actual class rate is over
Corey Hickey wrote:
Hello,
I haven't been keeping up with sending ESFQ [ANNOUNCE] messages to this
list, but I've still been working on the patch. If you're curious about
recent changes, take a look at the home page, ChangeLog, and README:
http://fatooh.org/esfq-2.6/
Corey Hickey wrote:
Patrick McHardy wrote:
Should ESFQ be merged into SFQ or remain as a separate qdisc?
I've CCed netdev. I think merging parts of ESFQ (dynamic depth and
flow number) would make a lot of sense, but I'm intending to submit
an alternative to the ESFQ hashing scheme
Please send bugreports to [EMAIL PROTECTED]
Ole Reinartz wrote:
I'm trying to get some DiffServ QoS shaping to work on an XScale
machine, running big endian. I'm setting it up with tc. Using the
tcindex filter I found that regardless what shift value I enter, only
'0' is returned when I list
ArcosCom Linux User wrote:
Any help please?
Please attach your scripts, your mailer wrapped the lines which
makes them pretty unreadable.
___
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
Thomas Graf wrote:
@@ -242,10 +239,10 @@ static int fib4_rule_compare(struct fib_
return 0;
#endif
- if (tb[FRA_SRC] (rule4-src != nla_get_be32(tb[FRA_SRC])))
+ if (frh-src_len (rule4-src != nla_get_be32(tb[FRA_SRC])))
return 0;
- if
Thomas Graf wrote:
* Patrick McHardy [EMAIL PROTECTED] 2007-03-20 17:59
The presence of the attributes when src_len/dst_len is non-zero
is only verified in fib_newrule, so this looks like it might crash
when something broken sets src_len/dst_len to a non-zero value
without actually adding
Thomas Graf wrote:
* Patrick McHardy [EMAIL PROTECTED] 2007-03-19 06:54
Thomas, I can't see a clean way to fix this right now that
doesn't either bloat struct nla_policy or removes FRA_SRC/FRA_DST
from the policy, could you please look into this? Thanks.
I guess the only way is to remove
Patrick McHardy wrote:
[NET]: Fix fib_rules compatibility breakage
I forgot to remove FRA_SRC/FRA_DST from fib6_rule_policy.
Updated patch attached.
[NET]: Fix fib_rules compatibility breakage
The fib_rules netlink attribute policy introduced in 2.6.19 broke
userspace compatibilty. When
Luciano Ruete wrote:
After an:
# ip ru flush
I loose all my ip rules but the priority 0 one.
[EMAIL PROTECTED]:~# ip ru
0: from all lookup 255
[EMAIL PROTECTED]:~#
Ok with that, but now i'm not able to insert any new rule.
This leads to a total loose of conectivity.
[EMAIL
Arkadiusz Miskiewicz wrote:
This patch
http://www.mail-archive.com/netdev@vger.kernel.org/msg27506.html
didn't make into upstream linux kernel it seems.
As mentioned in the changelog, its in 2.6.19.
The question is - are patches adding some functionality that's not in
upstream
kernel
Arkadiusz Miskiewicz wrote:
Hm, why no RTA_FWMASK in HEAD rtnetlink.h
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=include/linux/rtnetlink.h;h=4a629ea70cc4ca60a6f486f8653974af68dbe8cd;hb=HEAD
and 2.6.19
Stephen Hemminger wrote:
On Wed, 14 Mar 2007 13:32:28 +0100
Patrick McHardy [EMAIL PROTECTED] wrote:
RTA_* attributes aren't used for routing rules anymore inside
the kernel.
But we need to keep them in iproute2 for back compatibility?
Not really for compatibility, its the numerical
Krzysztof Oledzki wrote:
Getting back to the question: generally I have no objection for
forwarding connlinit to the mainline but I believe we should first
investigate a possibilty to add support for other protocols than TCP.
AFAIK at least UDP support could be very usefull - p2p software
Simon Lodal wrote:
This patch changes HTB's class storage from hash+lists to a two-level linear
array, so it can do constant time (O(1)) class lookup by classid. It improves
scalability for large number of classes.
Without the patch, ~14k htb classes can starve a Xeon-3.2 at only 15kpps,
Jan Engelhardt wrote:
On Jan 10 2007 06:58, Patrick McHardy wrote:
I would prefer to have someone maintain it externally though. Jan, are
you still interested in doing that? If you need help or webspace for
an external repository please let me know.
I would give it a try. Though I would
ArcosCom Linux User wrote:
The configuration is:
1) linux box with 2.6.19.1 kernel with these patches/modules:
a) l7-filter
b) multipath patch (from nano-howto)
c) IMQ
d) ipp2p
e) connlimit
2) 4 ethernet interfaces:
a) 2 external (eth1 and eth3)
Krzysztof Oledzki wrote:
Its still down, but the ROUTE patch is unmaintained anyway.
How about attached (and inlined) patch. BTW - is it possible to add a
Kconfig entry after a specific text, like with Makefile.ladd?
[POM-NG] ROUTE: 2.6.19 compatibility fix
Make both IPv4 and IPv6
ArcosCom Linux User wrote:
The log says:
Dec 30 00:52:27 cura kernel: dst cache overflow
Dec 30 00:52:27 cura kernel: MASQUERADE: No route: Rusty's brain broke!
Dec 30 00:52:27 cura kernel: dst cache overflow
Dec 30 00:52:28 cura kernel: zlan0: received tcn bpdu on port 1(eth0)
Dec 30
ArcosCom Linux User wrote:
Thanks for your response.
I'm using multiple gateways for internet connection and having problems
with random disconection, and I not use ROUTE usually, but I was trying to
force only one gateway for one type of traffic (which the clients lost
conections and are
ArcosCom Linux User wrote:
Then, the actual and updated and maintained substitute for ROUTE is using
CONNMARK and/or MARK and then add filters/rules to routes table with ip.
Am I in the truth?
That has always been the better way. The route target is a hack, I'm
don't know why it exists at
François Delawarde wrote:
I was thinking of trying that along with the netfilter SIP helper, but I
don't even understand how helpers work yet. If you have an idea of how i
could use those things, it would also be worth trying.
Just load ip_nat_sip, it should adjust the SDP information
Flechsenhaar, Jon J wrote:
$TC qdisc add dev $EDEV parent 2:20 gred setup DPs 3 default 2 grio
after each DP 3 on each gred.
This starting happening after I upgraded to 2.6.18 from 2.4.20 kernel.
Anyone have any ideas?´
I think DPs start at zero, so you have 0, 1 and 2. 3 is out of bounds
Thossapron Apinyapanha wrote:
tc command's HFSC have a lot parameter with 4 curve type
SC curve - umax dmax rate
LS curve - umax dmax rate
RT curve - umax dmax rate
UL curve - umax dmax rate
so i'dont know which parameter are appropriate for my test case
such real time class which curve
Thossapron Apinyapanha wrote:
after i use tc -s -d class ls dev eth0 will show statistic data about HFSC
,like this
[EMAIL PROTECTED]:/home/shaper# tc -s -d class ls dev eth2
class hfsc 1: root
Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
period 24 work 13844792199226589188
HFSC can operate in non-work-conserving mode itself (using upper-limit
curves), so there is no need to attach further non-work-conserving
qdiscs as leaves.
- Message d'origine
De : Patrick McHardy [EMAIL PROTECTED]
À : Leo Wetz [EMAIL PROTECTED]
Cc : lartc@mailman.ds9a.nl
Envoyé le
Leo Wetz wrote:
Hello,
I have finally managed to understand HFSC up to a level which allowed me to
create a QoS script which maintains low VoIP latency while running stuff
like eMule.
Unfortunately, HFSC seems to have a severe bug.
Why do I consider this as a bug defenitely?
Well, my
Justin Schoeman wrote:
I haven't been paying attention to this for a while, but now that I
download the latest patch-o-matic-ng, I see that most of the patches are
gone...
Anybody have an idea where I can download the 'extras' repository?
Specifically geoip.
We removed all patches that we
Andy Furniss wrote:
Marlon Dutra wrote:
lended: 150 borrowed: 5077 giants: 3986
tokens: -14728330 ctokens: -21365
Regardless 11678670/2711 1500 so specify your mtu next to every rate
and ceil.
I created the class with mtu 1500 and the result is above, same
behaviour. Afaik, the default
Hi Martin,
first of all: thanks alot for your efforts (and sorry for a bit of
silence from my side, I'm busy as usual ..)
Martin A. Brown wrote:
Greetings Nickola!
: Just a question - wasn't Mr. Kenjiro Cho [1] the original writer
: of the HFSC queueing discipline, following the work of
jamal wrote:
On Tue, 2006-20-06 at 02:54 +0200, Patrick McHardy wrote:
jamal wrote:
- For further reflection: Have you considered the case where the rate
table has already been considered on some link speed in user space and
then somewhere post-config the physical link speed changes
jamal wrote:
On Tue, 2006-20-06 at 03:04 +0200, Patrick McHardy wrote:
It would be nice to have support for HFSC as well, which unfortunately
needs to be done in the kernel since it doesn't use rate tables.
What about qdiscs like SFQ (which uses the packet size in quantum
calculations)? I guess
jamal wrote:
- For further reflection: Have you considered the case where the rate
table has already been considered on some link speed in user space and
then somewhere post-config the physical link speed changes? This would
happen in the case where ethernet AN is involved and the partner
jamal wrote:
You are still speaking ATM (and the above may still be valid), but:
Could you for example look at the netdevice-type and from that figure
out the link layer overhead and compensate for it.
Obviously a lot more useful if such activity is doable in user space
without any knowledge
Eliot, Wireless and Server Administrator, Great Lakes Internet wrote:
Eh. What a pain. If I disable this, then ebtables will not call iptables
after the ebtables are finished running. I figured out that I could use
ebtables to match the destination MAC address like I needed for the
other
Eliot, Wireless and Server Administrator, Great Lakes Internet wrote:
However, this still does not work:
Chain POSTROUTING (policy ACCEPT 812K packets, 441M bytes)
pkts bytes target prot opt in out source
destination
2071 129K CLASSIFY all -- * br1 0.0.0.0/0
Eliot, Wireless and Server Administrator, Great Lakes Internet wrote:
These rules make it go to the classes instead of the qdisc:
Chain POSTROUTING (policy ACCEPT 887K packets, 495M bytes)
pkts bytes target prot opt in out source
destination
8662 508K CLASSIFY all -- *
Eliot, Wireless and Server Administrator, Great Lakes Internet wrote:
Both devices (br1 and wivl4) are bridged interfaces with spanning tree
turned on. They also do VLANs. Specifically, vconfig was used to create
a VLAN (in this case, VLAN 4) on two interfaces: eth2 and eth3. These
two VLAN
Eliot, Wireless and Server Administrator, Great Lakes Internet wrote:
Bridged iptables (ebtables) is not enabled in the kernel and I cannot
seem to find a variable bridge-nf-call-iptables to set with sysctl:
wireless-r1 linux # sysctl -w bridge-nf-call-iptables=0
error:
Eliot, Wireless and Server Administrator, Great Lakes Internet wrote:
THANK YOU!
That solved the problem. I found the file you specified and it was
indeed enabled. After disabling it, it is now working!
Good to hear. This crap is causing one weird problem after another,
we really need to get
Alexandru Dragoi wrote:
I think i'd like more docs in english about hfsc.
Me too. I don't have time to write one myself (and I'm not good at
this), but I can assist if anyone wants to do it.
I would like to know
also some tips about scalability at large amount of traffic, like more
than
Jody Shumaker wrote:
HFSC doesn't support strict priorities (and neither does HTB, the
priorities just affect unused bandwidth and is still limited by the
ceiling). At least in the case of HFSC this is intentional, strict
priority is not very friendly because it allows traffic to be
entirely
G Georgiev wrote:
OK,
Found a solution - if some is interested - assigned the near end of
the IPSEC tunnel address to the internal interface; this way got a
POSTROUTING chain available and did an SNAT there:
ip addr add 10.253.0.2 dev eth0;
ip route add to unicast
Jody Shumaker wrote:
My understanding of HFSC is limited, but i'm fairly sure its similar
to all other qdiscs in one respect that would make the config you have
shown, not actually work as you've described. Each of those HFSC
qdiscs is a seperate entity, no sharing will occur between those
Patrick McHardy wrote:
Jody Shumaker wrote:
What is there for good HSFC documentation out there right now anyways?
There is the original papers by Hui Zhang et al., which is mostly
about the theory and not very suitable for users - but still worth
reading if you're not scared by use
Andreas Mueller wrote:
I allways forget attachments. ;)
--- linux/net/sched/sch_hfsc.c~ 2006-01-15 07:16:02.0 +0100
+++ linux/net/sched/sch_hfsc.c2006-05-10 00:07:07.0 +0200
@@ -970,14
G Georgiev wrote:
Hi,
Could not conceive an working set-up for an IPSEC VPN made with
racoon/setkey
on which I have one address on my side acting as an SNAT router for all
traffic from my network to a network segment on the far side.
my network --- my gateway
Jason Boxman wrote:
On Wednesday 22 March 2006 19:13, James Nelson wrote:
Thanks for all of your help Patrick!
Just so I'm clear. If hfsc at the class level shows no overlimits and no
packet dropps, then hfsc is not effecting my traffic any different (from a
throughput perspective
James Nelson wrote:
I don't understand the following:
[EMAIL PROTECTED]:~# tc -s class ls dev vlan1 tc -s qdisc ls dev vlan1
class hfsc 1: root
Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
period 0 level 2
class hfsc 1:1 parent 1: sc m1 0bit d 0us m2 22bit ul m1 0bit d 0us m2
Kirk Reiser wrote:
Patrick McHardy [EMAIL PROTECTED] writes:
Please post your kernel version, your iptables version and the
output of iptables -vxnL.
Woops! The kernel is linux 2.6.15.6 and the iptables is 1.3.3. I
will have to reconstruct the script using multiport so that will take
Jody Shumaker wrote:
Could whomever is in charge of the lartc mailing list please change it
to add the header:
Reply-To: lartc@mailman.ds9a.nl
Every other list I'm on is setup so that by default replies will go to
the list. When replying to lartc emails I notice myself and others
jamal wrote:
On Mon, 2006-13-03 at 14:44 +1000, Russell Stuart wrote:
You are wrong on both counts.
I am wrong on why it is being rejected - but what you are seeing is
worse than i thought initially.
Lets put it this way:
The only you will _ever_ get that message is if you had made
Stephen Hemminger wrote:
The memset fix is in current CVS. I just wasn't going to take the
patch that looked at utsname to decide what hash to use.
Yes, that sucks. We have another incompatibility, old iproute versions
(like the one shipped by Debian) show garbage statistics for HFSC. I
think
Stephen Hemminger wrote:
BTW, running valgrind on tc shows lots of uses of uninitialized values,
it seems like a good idea if someone would go over these and fix them
up.
If we had a test script of commands (code coverage), that would help.
Actually the Coverity scanner is quite good at
James Nelson wrote:
Hi,
I'm trying to get a handle on hfsc. Here is my configuration:
[EMAIL PROTECTED]:/jmn# tc class show dev vlan1
class hfsc 1: root
class hfsc 1:1 parent 1: ls m1 0bit d 0us m2 225000bit ul m1 0bit d 0us m2
225000bit
class hfsc 1:10 parent 1:1 rt m1 191000bit d
Patrick McHardy wrote:
James Nelson wrote:
[EMAIL PROTECTED]:/jmn# tc class show dev vlan1
[...]
Why is there dropped packets but nothing overlimits??
Overlimits counts dequeue-attempts that were unsuccessful because of
qdisc limits. Drops usually occur when the inner qdisc is full
Salim wrote:
Hi All,
I am adding ip_queue module for snort inline IDS.
I am using snort2.4.0
And iptables-1.3.4.
Userspace Queuing(queue target) is enabled. It is built-in and not built as
a module.
The output of /proc/net/ip_queue is shown below:
cat /proc/net/ip_queue
Peer PID :
Ethy H. Brito wrote:
I am kinda frustrated with the lack of help some developers are dispensing to
this
problem (read ignoring it).
You should report it to a list that it actually read by developers,
namely [EMAIL PROTECTED]
___
LARTC mailing list
Salim wrote:
it does work when iptables as a whole is built as a module.
Do you use any patches that might register as queue handler,
like IMQ? Otherwise please check your logs for messages from
ip_queue during boot time, it should have logged the reason
if registration failed.
Jones Desougi wrote:
Try the patch below. (It's bug #413 in bugzilla)
Applied, thanks Jones.
___
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
DervishD wrote:
Hi all :)
If I set the burst/cburst parameter to, let's say, 1500, the
command tc -s -d class show dev eth0 says that the value is 1499b/8
instead of the (correct?) 1500b/8.
Is this right or am I doing anything wrong?
No, this is related to an integer division
DervishD wrote:
Hi Salim :)
* Salim [EMAIL PROTECTED] dixit:
I got this problem while trying to shape traffic with iptables MARK and
HTB.
MARK: targinfosize 8 != 4
--set-mark gives invalid argument error message.
Kernel version is 2.4.29 (some patches from patch o matic applied)
Jones Desougi wrote:
That can't be the reason, all revisions of a single match/target are
in the same object file and the supported revision is (supposed to be)
probed. Salim, can you send a strace of the failing iptables command?
The key being supposed to be. :-)
I somehow expected
Brian J. Murrell wrote:
I thought I had this all worked out, but it seems not. The following tc
configuration:
tc qdisc del dev ppp0 root 2 /dev/null /dev/null
tc qdisc add dev ppp0 root handle 1: tbf rate 120kbit burst 1200 limit 1
But it seems that some outbound flows are being blocked
Andreas Unterkircher wrote:
Good suggestion to use ulog for this. So I could dump the exactly
traffic which would run through a class (CLASSIFY)
to analyze and extract the necessary data to draw the graphs. So I do
not have to parse my class (IP or MAC) out of a
full tcpdump stream.
Sadly not
Andreas Unterkircher wrote:
Hello list,
I'm currently a bit planless so perhaps someone here could give me a point in
the right direction.
History: I wrote a shaper web tool (http://shaper.netshadow.at) and now got
several feature requests if it would be possible to graph what's going on
(this
Chris Kloosterman wrote:
- We have two IP addresses assigned to this machine using aliases:
[EMAIL PROTECTED] ~]# ip addr show
2: eth0: BROADCAST,MULTICAST,UP mtu 1500 qdisc pfifo_fast qlen 1000
...
inet xxx.yyy.zzz.39/24 brd xxx.yyy.zzz.255 scope global eth0
inet xxx.yyy.zzz.16/24 brd
Damjan wrote:
Patrick McHardy
* Fix ip command shortcuts
Hmm.. what's this change?
I've noticed that ip address no longer works, only ip addr works.
This seems to be one of the things the introduction of
batch mode broke. This patch fixes it.
Index: ip/ip.c
Evgeny Kushmanov wrote:
When I do ip ru ls very often is nothing output, but sometime it
output correct info
[EMAIL PROTECTED]:~# ip ru ls
[EMAIL PROTECTED]:~# ip ru ls
[EMAIL PROTECTED]:~# ip ru ls
[EMAIL PROTECTED]:~# ip ru ls
[EMAIL PROTECTED]:~# ip ru ls
[EMAIL PROTECTED]:~# ip ru ls
Carl-Daniel Hailfinger wrote:
Patrick McHardy schrieb:
Why do you want to decrease speed as the quota is approached?
We have two phases (simplified):
1. Already sent traffic is less than htbq_squota
- Do not limit anything.
2. Already sent traffic is more than htbq_squota
Carl-Daniel Hailfinger wrote:
Proposal:
--
Another idea would be to create a qdisc HTBQ (HTB with quota)
derived from HTB with the following characteristics:
htb_rate=min(htbq_rate,
(alreadysent=htbq_squota)?((htbq_quota-alreadysent)/remtime):htbq_rate)
Forte Systems - Iosif Peterfi wrote:
The tc show command doesnt create any sent/rate statistics for
1: / 1:1 / 1:2 / 1:3 Everything is 0 except the period indicators.
Sent bytes are shown in the period, but the rate is not shown. For the
rest of the classes Sent bytes and packets are
Stephen Hemminger wrote:
There is an design problem with the qdisc interface that causes qlen related
bugs
in netem, tbf, and other qdisc's that peek at the top of the queue. The
problem is
that requeue needs to be called from the dequeue function but requeue can
fail.
If requeue fails,
Antonio Pinizzotto wrote:
Hi everybody.
Do you know about any way to read the TCP cwnd value (congestion window)
on Linux?
I have read that on Linux it is not possible to enable a socket option
(to read to cwnd using the program trpt).
Any way to read the cwnd would be good for me.
I
R Harper wrote:
TBF provides traffic shaping by the Token Bucket theory, while SFQ makes
sure(actually just hints) swap packets in different sessions so that no
particular session will hang around for a long time.
Yes I know the difference between TBF and SFQ.
I was trying to ask about the
Kunszt Arpad wrote:
I use the CPU as timer and I have a dual Xeon box. So I use the SMP but
I have phisically 2 CPUs. I don't use HyperThreading.
The TSCs on an SMP box may drift apart. Can you reproduce the problem
with gettimeofday as time source?
Regards
Patrick
Wang Jian wrote:
Hi,
One of my customer needs per flow rate control, so I write one.
The code I post here is not finished, but it seems to work as expected.
The kernel patch is agains kernel 2.6.11, the iproute2 patch is against
iproute2-2.6.11-050314.
I write the code in a hurry to meet
Tero Saarni wrote:
I'm trying to configure dsmark qdisc on 2.6.11.4 user mode linux and
tc from iproute2-2.6.11-050314.
I think I have some mismatch in my setup since adding dsmark qdisc
fails *unless* I specify set_tc_index argument which I believe should
be optional:
# tc qdisc add dev eth1
Andy Furniss wrote:
Seems still broken.
I built vanilla(apart from nth) 2.6.10, new iptables 1.2.11 +
pom-200400621 with runme extra only said y to nth.
I see -
Chain OUTPUT (policy ACCEPT 817 packets, 103K bytes)
pkts bytes target prot opt in out source destination
15 1260 MARK
Catalin(ux aka Dino) BOIE wrote:
Hello!
I am glad to announce a patch for u32 to allow matches on nfmark.
The patch is non intrusive (few lines).
if ((*(u32*)(ptr+key-off+(off2key-offmask))^key-val)key-mask) {
---
Michael Harris wrote:
When i try adding these rules:
ip route add default via X.X.X.1 dev eth0 table 200
ip route add nat X.X.X.6 via Z.Z.Z.46
ip rule add from Z.Z.Z.46 nat X.X.X.6 table 200
the second route causes this error: RTNETLINK answers: File exists
i have also tried changing the command
Vincent Perrier wrote:
HTB versus HFSC, both qdisc offer the same kind of service,
if you want to see comparative test results, go to
http://www.rawsoft.org
at the line TEST RESULTS you will find the results for
a sharing test and a burst test.
You will see that both qdisc are good.
Nice
Dmitry Golubev wrote:
Quote from LATRC: Also, with HTB, you should attach all filters to the root!
Then LARTC is wrong. HTB first tries to classify by priority,
if that fails it follows the filter chain from the root until
it hits a leaf-class, if that fails it tries if the default
class is a
Andy Furniss wrote:
Jiri Fojtasek wrote:
It leave always only one packet in the device queue. It is there until
is not dequeued (sent to the device driver) and then is inserted
another etc, etc
I think I understand - in the case of imq - ppp - usb/pci - radsl
modem with big buffer - phone
Nuutti Kotivuori wrote:
But, as the implementation is rather old, it would require a complete
overhaul for 2.6 I think. It is a shame it wasn't worked in to the
Linux kernel when it was still current, as I think the algorithm could
have a lot of uses.
I've completed the port and tested it
Nuutti Kotivuori wrote:
Patrick McHardy wrote:
I've completed the port and tested it yesterday, unfortunately it's
not useable in the real world as is. There is a strong bias against
non-ECN flows because their packets are simply dropped instead of
marked. At high load (50 ECN vs. 1 non-ECN flow
Nuutti Kotivuori wrote:
Patrick McHardy wrote:
There is a blue implementation for Linux at
http://home.sch.bme.hu/~bartoki/projects/thesis/2001-May-26/
Nice!
I briefly scanned the implementation and it didn't look too bad. Some
oddities here and there (such as changing HZ from 100 to 1024
Nuutti Kotivuori wrote:
Patrick McHardy wrote:
This is currently all there is. If you have some specific questions,
just ask (but please CC lartc). If anyone wants to write some
documentation I'd be happy to help, but I don't have time for it
myself.
I am not sure if the original poster has
Andre,
thanks for doing this. I also felt like I need to say something for
some time. Roy, you should also stop telling people your version would
be more stable. IMQ's problems are related to specific setups, so I
don't understand how you can state this without even understanding the
problem.
Santiago J. Ruano Rincón wrote:
try HFSC, Hierarchical Fair Service Curve:
http://trash.net/~kaber/hfsc
http://www-2.cs.cmu.edu/~hzhang/HFSC/
i'm going to test it in this week.
This should indeed work with HFSC if the custumer-classes have only
links-haring service curves and no realtime
Vincent Jaussaud wrote:
Hi;
First, sorry if this question is mostly netfilter related, than lartc,
but I think you guys may have a your opinion about this.
You should post this question to the netfilter-devel list, people there
can give you very good advice.
I'm using Linux 2.4.x with netfilter
Andre Correa wrote:
Patrick, tks for the info but I'm sure I got your idea.
A filter handle is something like: 804::800 right?
Not exactly. How handles are handled depends on the classifier,
fw classifier for example uses its own handle to match the nfmark,
route creates handles of its own and
Andre Correa wrote:
Hi list, I'm playing with tc and found a strange behavior when I try
to delete filters. For example, this simple scenario:
tc qdisc add dev eth1 root handle 1: htb default 100
tc class add dev eth1 parent 1: classid 1:1 htb rate 128Kbit
tc class add dev eth1 parent 1:
[EMAIL PROTECTED] wrote:
Hi,
While going through the README of Michal's SNMP extension, he collects stats using netlink interface. Is their any howto or other docs to learn how it works?
raj
___
LARTC mailing list / [EMAIL PROTECTED]
Tilman Giese (Global View) wrote:
Hi,
I am experiencing a curious phenomenon. I limited the bandwidth for a
specific client to 750KBit. It works well despite of the fact that the
client always gets a little bit more bandwidth (around 770KBit to
780KBit). I used different bandwidth and traffic
1 - 100 of 152 matches
Mail list logo
