* [HELP] for using class group classifier network controller
@ 2009-06-15 0:56 ` 김재열
0 siblings, 0 replies; 12+ messages in thread
From: 김재열 @ 2009-06-15 0:56 UTC (permalink / raw)
To: usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil, tgraf-wEGCiKHe2LqWVfeAwA7xHQ
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Hi.
I'm trying to use class group classifier network controller written by Thomas Graf.
But it doesn't work well.
I need a help to use this network classifier collectly
Below is the my testing environment
kernel : 2.6.29
iproute : iproute2-2.6.29-1 + net_cls patch
I compiled kernel in order to use control group classifier by enabling CGROUPS and NET_CLS_CGROUPS and other cgroup related options.
And I operated like this.
mkdir -p /dev/cpuctl/
mount -t cgroup net_cls -onet_cls /dev/cpuctl
mkdir /dev/cpuctl/foo
mkdir /dev/cpuctl/bob
# classid 1:10
echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
#classid 1:20
echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
And I put two bashshell's pid into foo and bob group
echo $$ > /dev/cpuctl/foo/task
echo $$ > /dev/cpuctl/bob/task
And then I configured the kernel's traffic control policy
TC = /sbin/tc
$TC qdisc delete dev eth0 root
$TC qdisc add dev eth0 root handle 1: htb default 10 r2q 200
# add class 1:10 for foo group
$TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
# add class 1:20 for bob group
$TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
# default class
$TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
# add net_cls filter for class foo and bob
$TC filter add dev eth0 parent 1: protocol ip prio 10 handle 100: cgroup
# default class filter
$TC filter add dev eth0 parent 1: protocol ip prio 20 basic classid 1:30
After finishing this configuration, I ran the iperf for measuring the network BW.
Iperf showd about 10mbps in each group.
10mbps is the rate of class 1:30, default class.
When I send a data in group foo or group bob, the data rate is equal to default class 1:30 rate,
so I think kernel couldn't identify classified packet from normal packet.
What is wrong?
I expect your kindness.
- Cheiyol Kim
*************************************************************************************
This is the iproute patch
--- ./iproute2-2.6.29-1/tc/Makefile 2009-03-25 07:40:54.000000000 +0900
+++ ./iproute2-2.6.29-1-mod/tc/Makefile 2009-06-02 17:33:21.000000000 +0900
@@ -20,6 +20,7 @@
TCMODULES += f_fw.o
TCMODULES += f_basic.o
TCMODULES += f_flow.o
+TCMODULES += f_cgroup.o
TCMODULES += q_dsmark.o
TCMODULES += q_gred.o
TCMODULES += f_tcindex.o
--- ./iproute2-2.6.29-1/tc/NULL 1970-01-01 09:00:00.000000000 +0900
+++ ./iproute2-2.6.29-1-mod/tc/f_cgroup.c 2009-06-02 17:32:47.000000000 +0900
@@ -0,0 +1,101 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include "utils.h"
+#include "tc_util.h"
+#include "m_ematch.h"
+
+static void explain(void)
+{
+ fprintf(stderr, "Usage: ... cgroup [ match EMATCH_TREE ] [ police POLICE_SPEC ]\n");
+ fprintf(stderr, " [ action ACTION_SPEC ]\n");
+}
+
+static int cgroup_parse_opt(struct filter_util *qu, char *handle,
+ int argc, char **argv, struct nlmsghdr *n) {
+ struct tcmsg *t = NLMSG_DATA(n);
+ struct rtattr *tail;
+ long h = 0;
+
+ if (handle) {
+ h = strtol(handle, NULL, 0);
+ if (h == LONG_MIN || h == LONG_MAX) {
+ fprintf(stderr, "Illegal handle \"%s\", must be numeric.\n",
+ handle);
+ return -1;
+ }
+ }
+
+ t->tcm_handle = h;
+
+ tail = (struct rtattr*)(((void*)n)+NLMSG_ALIGN(n->nlmsg_len));
+ addattr_l(n, MAX_MSG, TCA_OPTIONS, NULL, 0);
+
+ while (argc > 0) {
+ if (matches(*argv, "match") == 0) {
+ NEXT_ARG();
+ if (parse_ematch(&argc, &argv, TCA_CGROUP_EMATCHES, n)) {
+ fprintf(stderr, "Illegal \"ematch\"\n");
+ return -1;
+ }
+ continue;
+ } else if (matches(*argv, "action") == 0) {
+ NEXT_ARG();
+ if (parse_action(&argc, &argv, TCA_CGROUP_ACT, n)) {
+ fprintf(stderr, "Illegal \"action\"\n");
+ return -1;
+ }
+ continue;
+
+ } else if (matches(*argv, "police") == 0) {
+ NEXT_ARG();
+ if (parse_police(&argc, &argv, TCA_CGROUP_POLICE, n)) {
+ fprintf(stderr, "Illegal \"police\"\n");
+ return -1;
+ }
+ continue;
+ } else if (strcmp(*argv, "help") == 0) {
+ explain();
+ return -1;
+ } else {
+ fprintf(stderr, "What is \"%s\"?\n", *argv);
+ explain();
+ return -1;
+ }
+ argc--; argv++;
+ }
+
+ tail->rta_len = (((void*)n)+n->nlmsg_len) - (void*)tail;
+ return 0;
+}
+
+static int cgroup_print_opt(struct filter_util *qu, FILE *f,
+ struct rtattr *opt, __u32 handle) {
+ struct rtattr *tb[TCA_CGROUP_MAX+1];
+
+ if (opt == NULL)
+ return 0;
+
+ parse_rtattr_nested(tb, TCA_CGROUP_MAX, opt);
+
+ if (handle)
+ fprintf(f, "handle 0x%x ", handle);
+
+ if (tb[TCA_CGROUP_EMATCHES])
+ print_ematch(f, tb[TCA_CGROUP_EMATCHES]);
+
+ if (tb[TCA_CGROUP_POLICE]) {
+ fprintf(f, "\n");
+ tc_print_police(f, tb[TCA_CGROUP_POLICE]);
+ }
+
+ if (tb[TCA_CGROUP_ACT])
+ tc_print_action(f, tb[TCA_CGROUP_ACT]);
+
+ return 0;
+}
+
+struct filter_util cgroup_filter_util = {
+ .id = "cgroup",
+ .parse_fopt = cgroup_parse_opt,
+ .print_fopt = cgroup_print_opt,
+};
^ permalink raw reply [flat|nested] 12+ messages in thread
* [HELP] for using class group classifier network controller
@ 2009-06-15 0:56 ` 김재열
0 siblings, 0 replies; 12+ messages in thread
From: 김재열 @ 2009-06-15 0:56 UTC (permalink / raw)
To: usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil, tgraf-wEGCiKHe2LqWVfeAwA7xHQ
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Hi.
I'm trying to use class group classifier network controller written by Thomas Graf.
But it doesn't work well.
I need a help to use this network classifier collectly
Below is the my testing environment
kernel : 2.6.29
iproute : iproute2-2.6.29-1 + net_cls patch
I compiled kernel in order to use control group classifier by enabling CGROUPS and NET_CLS_CGROUPS and other cgroup related options.
And I operated like this.
mkdir -p /dev/cpuctl/
mount -t cgroup net_cls -onet_cls /dev/cpuctl
mkdir /dev/cpuctl/foo
mkdir /dev/cpuctl/bob
# classid 1:10
echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
#classid 1:20
echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
And I put two bashshell's pid into foo and bob group
echo $$ > /dev/cpuctl/foo/task
echo $$ > /dev/cpuctl/bob/task
And then I configured the kernel's traffic control policy
TC = /sbin/tc
$TC qdisc delete dev eth0 root
$TC qdisc add dev eth0 root handle 1: htb default 10 r2q 200
# add class 1:10 for foo group
$TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
# add class 1:20 for bob group
$TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
# default class
$TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
# add net_cls filter for class foo and bob
$TC filter add dev eth0 parent 1: protocol ip prio 10 handle 100: cgroup
# default class filter
$TC filter add dev eth0 parent 1: protocol ip prio 20 basic classid 1:30
After finishing this configuration, I ran the iperf for measuring the network BW.
Iperf showd about 10mbps in each group.
10mbps is the rate of class 1:30, default class.
When I send a data in group foo or group bob, the data rate is equal to default class 1:30 rate,
so I think kernel couldn't identify classified packet from normal packet.
What is wrong?
I expect your kindness.
- Cheiyol Kim
*************************************************************************************
This is the iproute patch
--- ./iproute2-2.6.29-1/tc/Makefile 2009-03-25 07:40:54.000000000 +0900
+++ ./iproute2-2.6.29-1-mod/tc/Makefile 2009-06-02 17:33:21.000000000 +0900
@@ -20,6 +20,7 @@
TCMODULES += f_fw.o
TCMODULES += f_basic.o
TCMODULES += f_flow.o
+TCMODULES += f_cgroup.o
TCMODULES += q_dsmark.o
TCMODULES += q_gred.o
TCMODULES += f_tcindex.o
--- ./iproute2-2.6.29-1/tc/NULL 1970-01-01 09:00:00.000000000 +0900
+++ ./iproute2-2.6.29-1-mod/tc/f_cgroup.c 2009-06-02 17:32:47.000000000 +0900
@@ -0,0 +1,101 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include "utils.h"
+#include "tc_util.h"
+#include "m_ematch.h"
+
+static void explain(void)
+{
+ fprintf(stderr, "Usage: ... cgroup [ match EMATCH_TREE ] [ police POLICE_SPEC ]\n");
+ fprintf(stderr, " [ action ACTION_SPEC ]\n");
+}
+
+static int cgroup_parse_opt(struct filter_util *qu, char *handle,
+ int argc, char **argv, struct nlmsghdr *n) {
+ struct tcmsg *t = NLMSG_DATA(n);
+ struct rtattr *tail;
+ long h = 0;
+
+ if (handle) {
+ h = strtol(handle, NULL, 0);
+ if (h == LONG_MIN || h == LONG_MAX) {
+ fprintf(stderr, "Illegal handle \"%s\", must be numeric.\n",
+ handle);
+ return -1;
+ }
+ }
+
+ t->tcm_handle = h;
+
+ tail = (struct rtattr*)(((void*)n)+NLMSG_ALIGN(n->nlmsg_len));
+ addattr_l(n, MAX_MSG, TCA_OPTIONS, NULL, 0);
+
+ while (argc > 0) {
+ if (matches(*argv, "match") == 0) {
+ NEXT_ARG();
+ if (parse_ematch(&argc, &argv, TCA_CGROUP_EMATCHES, n)) {
+ fprintf(stderr, "Illegal \"ematch\"\n");
+ return -1;
+ }
+ continue;
+ } else if (matches(*argv, "action") == 0) {
+ NEXT_ARG();
+ if (parse_action(&argc, &argv, TCA_CGROUP_ACT, n)) {
+ fprintf(stderr, "Illegal \"action\"\n");
+ return -1;
+ }
+ continue;
+
+ } else if (matches(*argv, "police") == 0) {
+ NEXT_ARG();
+ if (parse_police(&argc, &argv, TCA_CGROUP_POLICE, n)) {
+ fprintf(stderr, "Illegal \"police\"\n");
+ return -1;
+ }
+ continue;
+ } else if (strcmp(*argv, "help") == 0) {
+ explain();
+ return -1;
+ } else {
+ fprintf(stderr, "What is \"%s\"?\n", *argv);
+ explain();
+ return -1;
+ }
+ argc--; argv++;
+ }
+
+ tail->rta_len = (((void*)n)+n->nlmsg_len) - (void*)tail;
+ return 0;
+}
+
+static int cgroup_print_opt(struct filter_util *qu, FILE *f,
+ struct rtattr *opt, __u32 handle) {
+ struct rtattr *tb[TCA_CGROUP_MAX+1];
+
+ if (opt == NULL)
+ return 0;
+
+ parse_rtattr_nested(tb, TCA_CGROUP_MAX, opt);
+
+ if (handle)
+ fprintf(f, "handle 0x%x ", handle);
+
+ if (tb[TCA_CGROUP_EMATCHES])
+ print_ematch(f, tb[TCA_CGROUP_EMATCHES]);
+
+ if (tb[TCA_CGROUP_POLICE]) {
+ fprintf(f, "\n");
+ tc_print_police(f, tb[TCA_CGROUP_POLICE]);
+ }
+
+ if (tb[TCA_CGROUP_ACT])
+ tc_print_action(f, tb[TCA_CGROUP_ACT]);
+
+ return 0;
+}
+
+struct filter_util cgroup_filter_util = {
+ .id = "cgroup",
+ .parse_fopt = cgroup_parse_opt,
+ .print_fopt = cgroup_print_opt,
+};
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
2009-06-15 0:56 ` 김재열
(?)
(?)
@ 2009-06-15 2:36 ` Minoru Usui
-1 siblings, 0 replies; 12+ messages in thread
From: Minoru Usui @ 2009-06-15 2:36 UTC (permalink / raw)
To: 김재열
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
tgraf-wEGCiKHe2LqWVfeAwA7xHQ
Hi, Cheiyol
> # classid 1:10
>
> echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
>
> #classid 1:20
>
> echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
(snip)
> # add class 1:10 for foo group
>
> $TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
>
> # add class 1:20 for bob group
>
> $TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
>
> # default class
>
> $TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
In class ID to use TC, both a major number and a minor number are hexadecimal.
So 1:10 equals 0x00010010.
I think, you should set 0x10010 to net_cls.classid if you classify to 1:10.
--
Minoru Usui <usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil@public.gmane.org>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
2009-06-15 0:56 ` 김재열
(?)
@ 2009-06-15 2:36 ` Minoru Usui
2009-06-15 10:30 ` Thomas Graf
[not found] ` <20090615113609.db5c0613.usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil@public.gmane.org>
-1 siblings, 2 replies; 12+ messages in thread
From: Minoru Usui @ 2009-06-15 2:36 UTC (permalink / raw)
To: 김재열; +Cc: tgraf, netdev, containers
Hi, Cheiyol
> # classid 1:10
>
> echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
>
> #classid 1:20
>
> echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
(snip)
> # add class 1:10 for foo group
>
> $TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
>
> # add class 1:20 for bob group
>
> $TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
>
> # default class
>
> $TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
In class ID to use TC, both a major number and a minor number are hexadecimal.
So 1:10 equals 0x00010010.
I think, you should set 0x10010 to net_cls.classid if you classify to 1:10.
--
Minoru Usui <usui@mxm.nes.nec.co.jp>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
[not found] ` <20090615113609.db5c0613.usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil@public.gmane.org>
@ 2009-06-15 10:30 ` Thomas Graf
0 siblings, 0 replies; 12+ messages in thread
From: Thomas Graf @ 2009-06-15 10:30 UTC (permalink / raw)
To: Minoru Usui
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
On Mon, Jun 15, 2009 at 11:36:09AM +0900, Minoru Usui wrote:
> Hi, Cheiyol
>
> > # classid 1:10
> >
> > echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
> >
> > #classid 1:20
> >
> > echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
>
> (snip)
>
> > # add class 1:10 for foo group
> >
> > $TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
> >
> > # add class 1:20 for bob group
> >
> > $TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
> >
> > # default class
> >
> > $TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
>
> In class ID to use TC, both a major number and a minor number are hexadecimal.
> So 1:10 equals 0x00010010.
>
> I think, you should set 0x10010 to net_cls.classid if you classify to 1:10.
That's right, both net_cls.classid and tc use integers in hexadecimal form
for both, printing and parsing.
10:10 -> 0x100010
10:20 -> 0x100020
This seems to be the weak point regarding usability that's why I like
Minoru Usui's patch so much. It would no longer be necessary to really
understand the internals of classids etc.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
2009-06-15 2:36 ` Minoru Usui
@ 2009-06-15 10:30 ` Thomas Graf
[not found] ` <20090615103020.GA8894-PfSpb0PWhxZc2C7mugBRk2EX/6BAtgUQ@public.gmane.org>
2009-06-22 18:05 ` Paul Menage
[not found] ` <20090615113609.db5c0613.usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil@public.gmane.org>
1 sibling, 2 replies; 12+ messages in thread
From: Thomas Graf @ 2009-06-15 10:30 UTC (permalink / raw)
To: Minoru Usui; +Cc: 김재열, netdev, containers
On Mon, Jun 15, 2009 at 11:36:09AM +0900, Minoru Usui wrote:
> Hi, Cheiyol
>
> > # classid 1:10
> >
> > echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
> >
> > #classid 1:20
> >
> > echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
>
> (snip)
>
> > # add class 1:10 for foo group
> >
> > $TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
> >
> > # add class 1:20 for bob group
> >
> > $TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
> >
> > # default class
> >
> > $TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
>
> In class ID to use TC, both a major number and a minor number are hexadecimal.
> So 1:10 equals 0x00010010.
>
> I think, you should set 0x10010 to net_cls.classid if you classify to 1:10.
That's right, both net_cls.classid and tc use integers in hexadecimal form
for both, printing and parsing.
10:10 -> 0x100010
10:20 -> 0x100020
This seems to be the weak point regarding usability that's why I like
Minoru Usui's patch so much. It would no longer be necessary to really
understand the internals of classids etc.
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [Devel] Re: [HELP] for using class group classifier network controller
[not found] ` <20090615103020.GA8894-PfSpb0PWhxZc2C7mugBRk2EX/6BAtgUQ@public.gmane.org>
@ 2009-06-22 18:05 ` Paul Menage
0 siblings, 0 replies; 12+ messages in thread
From: Paul Menage @ 2009-06-22 18:05 UTC (permalink / raw)
To: Minoru Usui, 김재열,
netdev-u79uwXL29TY76Z2rM5mHXA,
containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
On Mon, Jun 15, 2009 at 3:30 AM, Thomas Graf<tgraf-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org> wrote:
>
> That's right, both net_cls.classid and tc use integers in hexadecimal form
> for both, printing and parsing.
net_cls.classid accepts integers in hexadecimal or decimal, and
reports in decimal.
Paul
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [Devel] Re: [HELP] for using class group classifier network controller
2009-06-15 10:30 ` Thomas Graf
[not found] ` <20090615103020.GA8894-PfSpb0PWhxZc2C7mugBRk2EX/6BAtgUQ@public.gmane.org>
@ 2009-06-22 18:05 ` Paul Menage
1 sibling, 0 replies; 12+ messages in thread
From: Paul Menage @ 2009-06-22 18:05 UTC (permalink / raw)
To: Minoru Usui, 김재열, netdev, containers
On Mon, Jun 15, 2009 at 3:30 AM, Thomas Graf<tgraf@infradead.org> wrote:
>
> That's right, both net_cls.classid and tc use integers in hexadecimal form
> for both, printing and parsing.
net_cls.classid accepts integers in hexadecimal or decimal, and
reports in decimal.
Paul
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
[not found] ` <d89801c9ef32$cdfcf480$8310fe81-tRM8rG9f+DVhl2p70BpVqQ@public.gmane.org>
@ 2009-06-18 4:13 ` Minoru Usui
0 siblings, 0 replies; 12+ messages in thread
From: Minoru Usui @ 2009-06-18 4:13 UTC (permalink / raw)
To: Chei-Yol Kim
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
tgraf-wEGCiKHe2LqWVfeAwA7xHQ
> when I set the classid to 0x10010, class 1:10's packet count was increased.
> But, according to your configuration, every packet has to flowed to 1:10 class.
Yes, you are right.
> Of course I know that the net_cls can't classify all packet because of the getting classid from the current task.
> Even though we consider it, Many packets are flowed to 1:30 class.
> therefore, the rate of the socket showed about 30mbps. This is the 1:30 class's rate.
> I think it's because that 1:30 class is the default class.
> When I changed the default to 20, the rate was varied to 20mbps, class 1:20's rate.
>
>
> To measure this rate, I used the iperf and scp command. Two results are the same.
> How about your experiment's result?
> Would you check the accurate rate of the network using applications like a iperf?
I did same experience with nuttcp, I observed the same result, too.
(Almost packets were classified into default classid.)
> By the results of my experiments, net_cls couldn't control the network rate of the group.
> What is your opinion?
I agree with you.
I'm going to investigate this problem.
--
Minoru Usui <usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil@public.gmane.org>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
[not found] <d89801c9ef32$cdfcf480$8310fe81@etri.info>
[not found] ` <d89801c9ef32$cdfcf480$8310fe81-tRM8rG9f+DVhl2p70BpVqQ@public.gmane.org>
@ 2009-06-18 4:13 ` Minoru Usui
1 sibling, 0 replies; 12+ messages in thread
From: Minoru Usui @ 2009-06-18 4:13 UTC (permalink / raw)
To: Chei-Yol Kim; +Cc: netdev, containers, tgraf
> when I set the classid to 0x10010, class 1:10's packet count was increased.
> But, according to your configuration, every packet has to flowed to 1:10 class.
Yes, you are right.
> Of course I know that the net_cls can't classify all packet because of the getting classid from the current task.
> Even though we consider it, Many packets are flowed to 1:30 class.
> therefore, the rate of the socket showed about 30mbps. This is the 1:30 class's rate.
> I think it's because that 1:30 class is the default class.
> When I changed the default to 20, the rate was varied to 20mbps, class 1:20's rate.
>
>
> To measure this rate, I used the iperf and scp command. Two results are the same.
> How about your experiment's result?
> Would you check the accurate rate of the network using applications like a iperf?
I did same experience with nuttcp, I observed the same result, too.
(Almost packets were classified into default classid.)
> By the results of my experiments, net_cls couldn't control the network rate of the group.
> What is your opinion?
I agree with you.
I'm going to investigate this problem.
--
Minoru Usui <usui@mxm.nes.nec.co.jp>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
[not found] ` <9f8101c9ed80$b730b830$8310fe81-tRM8rG9f+DVhl2p70BpVqQ@public.gmane.org>
@ 2009-06-15 8:18 ` Minoru Usui
0 siblings, 0 replies; 12+ messages in thread
From: Minoru Usui @ 2009-06-15 8:18 UTC (permalink / raw)
To: Chei-Yol Kim
Cc: netdev-u79uwXL29TY76Z2rM5mHXA,
containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA,
tgraf-wEGCiKHe2LqWVfeAwA7xHQ
On Mon, 15 Jun 2009 15:15:44 +0900
"Chei-Yol Kim" <gauri-mINagvnPBMnWF1p7bDn8Kw@public.gmane.org> wrote:
>
>
> >Hi, Cheiyol
> >
> >> # classid 1:10
> >>
> >> echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
> >>
> >> #classid 1:20
> >>
> >> echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
> >
> >(snip)
> >
> >> # add class 1:10 for foo group
> >>
> >> $TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
> >>
> >> # add class 1:20 for bob group
> >>
> >> $TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
> >>
> >> # default class
> >>
> >> $TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
>
> >In class ID to use TC, both a major number and a minor number are hexadecimal.
> >So 1:10 equals 0x00010010.
> >
> >I think, you should set 0x10010 to net_cls.classid if you classify to 1:10.
>
>
> I have done this by your advice. But it didn't work.
> I followed the instructions by http://thread.gmane.org/gmane.linux.network/108551/focus=108661
> So I supposed classid 1:10 is the decimal number, so I set the 0xA0001 as 1:10.
>
> I'm curious if this patch net_cls operates well in other's hands.
> How about you?
> Does your system using net_cls operate well?
> If then, would you let me know the environment and configuration of your system.
> I would reference your environment for solving my problem.
Which is your problem?
a) network packet can't be classified into the class which is set to net_cls.classid
b) network packet can be classified into the class which is set to net_cls.classid,
but its bandwidth limit doesn't work correctly.
If your problem is b), unfortunately I can't help you.
Because I am a begginer of cls_cgroup, and I haven't tested enough.
(But I think I might be facing same problem.)
If your problem is a), you should check 'tc -s filter show'.
You can see the result of classifing by cls_cgroup.
I can classified to the class which is set to root's net_cls.classid.
Environment:
kernel: 2.6.30
tc: iproute2-2.6.29-1 with tc/f_cgroup.c patch in latest git.
(I think it's same, probably)
How to use:
1) set to qdisc, class, filter
# tc qdisc add dev bond0 root handle 1: htb default 30
# tc class add dev bond0 parent 1:0 classid 1:10 htb rate 1mbit
# tc class add dev bond0 parent 1:0 classid 1:20 htb rate 2mbit
# tc class add dev bond0 parent 1:0 classid 1:30 htb rate 3mbit
# tc filter add dev bond0 protocol ip parent 1: prio 1 handle 0x1 cgroup
# tc -s class show dev bond0
class htb 1:10 root prio 0 rate 1000Kbit ceil 1000Kbit burst 1724b cburst 1724b
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0) <- packet count is 0
rate 0bit 0pps backlog 0b 0p requeues 0
lended: 0 borrowed: 0 giants: 0
tokens: 13476 ctokens: 13476
class htb 1:20 root prio 0 rate 2000Kbit ceil 2000Kbit burst 1849b cburst 1849b
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
lended: 0 borrowed: 0 giants: 0
tokens: 7226 ctokens: 7226
class htb 1:30 root prio 0 rate 3000Kbit ceil 3000Kbit burst 1974b cburst 1974b
Sent 25804 bytes 214 pkt (dropped 0, overlimits 0 requeues 0)
rate 3992bit 4pps backlog 0b 0p requeues 0
lended: 214 borrowed: 0 giants: 0
tokens: 4851 ctokens: 4851
2) mount cls_cgroup and set root's net_cls.classid
# mount -t cgroup -onet_cls none /cgroup/net/
# cd /cgroup/net/
# ls
net_cls.classid notify_on_release release_agent tasks
# echo 0x10010 > net_cls.classid # I want to classify 1:10
# cat net_cls.classid
65552 # 65552 = 65536 + 16 = 0x10010
3) using network
4) check the packet which is classified to the class which is set to net_cls.classid.
# /home/usui/src/iproute2-2.6.29-1/tc/tc -s class show dev bond0
class htb 1:10 root prio 0 rate 1000Kbit ceil 1000Kbit burst 1724b cburst 1724b
Sent 11238 bytes 95 pkt (dropped 0, overlimits 0 requeues 0) <- increase packet count
rate 2944bit 3pps backlog 0b 0p requeues 0
lended: 95 borrowed: 0 giants: 0
tokens: 12601 ctokens: 12601
class htb 1:20 root prio 0 rate 2000Kbit ceil 2000Kbit burst 1849b cburst 1849b
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
lended: 0 borrowed: 0 giants: 0
tokens: 7226 ctokens: 7226
class htb 1:30 root prio 0 rate 3000Kbit ceil 3000Kbit burst 1974b cburst 1974b
Sent 46092 bytes 342 pkt (dropped 0, overlimits 0 requeues 0)
rate 320bit 0pps backlog 0b 0p requeues 0
lended: 342 borrowed: 0 giants: 0
tokens: 5017 ctokens: 5017
--
Minoru Usui <usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil@public.gmane.org>
^ permalink raw reply [flat|nested] 12+ messages in thread
* Re: [HELP] for using class group classifier network controller
[not found] <9f8101c9ed80$b730b830$8310fe81@etri.info>
[not found] ` <9f8101c9ed80$b730b830$8310fe81-tRM8rG9f+DVhl2p70BpVqQ@public.gmane.org>
@ 2009-06-15 8:18 ` Minoru Usui
1 sibling, 0 replies; 12+ messages in thread
From: Minoru Usui @ 2009-06-15 8:18 UTC (permalink / raw)
To: Chei-Yol Kim; +Cc: tgraf, netdev, containers
On Mon, 15 Jun 2009 15:15:44 +0900
"Chei-Yol Kim" <gauri@etri.re.kr> wrote:
>
>
> >Hi, Cheiyol
> >
> >> # classid 1:10
> >>
> >> echo 0x1000A > /dev/cpuctl/foo/net_cls.classid
> >>
> >> #classid 1:20
> >>
> >> echo 0x10014 > /dev/cpuctl/bob/net_cls.classid
> >
> >(snip)
> >
> >> # add class 1:10 for foo group
> >>
> >> $TC class add dev eth0 parent 1:0 classid 1:10 htb rate 50mbit
> >>
> >> # add class 1:20 for bob group
> >>
> >> $TC class add dev eth0 parent 1:0 classid 1:20 htb rate 30mbit
> >>
> >> # default class
> >>
> >> $TC class add dev eth0 parent 1:0 classid 1:30 htb rate 10mbit
>
> >In class ID to use TC, both a major number and a minor number are hexadecimal.
> >So 1:10 equals 0x00010010.
> >
> >I think, you should set 0x10010 to net_cls.classid if you classify to 1:10.
>
>
> I have done this by your advice. But it didn't work.
> I followed the instructions by http://thread.gmane.org/gmane.linux.network/108551/focus=108661
> So I supposed classid 1:10 is the decimal number, so I set the 0xA0001 as 1:10.
>
> I'm curious if this patch net_cls operates well in other's hands.
> How about you?
> Does your system using net_cls operate well?
> If then, would you let me know the environment and configuration of your system.
> I would reference your environment for solving my problem.
Which is your problem?
a) network packet can't be classified into the class which is set to net_cls.classid
b) network packet can be classified into the class which is set to net_cls.classid,
but its bandwidth limit doesn't work correctly.
If your problem is b), unfortunately I can't help you.
Because I am a begginer of cls_cgroup, and I haven't tested enough.
(But I think I might be facing same problem.)
If your problem is a), you should check 'tc -s filter show'.
You can see the result of classifing by cls_cgroup.
I can classified to the class which is set to root's net_cls.classid.
Environment:
kernel: 2.6.30
tc: iproute2-2.6.29-1 with tc/f_cgroup.c patch in latest git.
(I think it's same, probably)
How to use:
1) set to qdisc, class, filter
# tc qdisc add dev bond0 root handle 1: htb default 30
# tc class add dev bond0 parent 1:0 classid 1:10 htb rate 1mbit
# tc class add dev bond0 parent 1:0 classid 1:20 htb rate 2mbit
# tc class add dev bond0 parent 1:0 classid 1:30 htb rate 3mbit
# tc filter add dev bond0 protocol ip parent 1: prio 1 handle 0x1 cgroup
# tc -s class show dev bond0
class htb 1:10 root prio 0 rate 1000Kbit ceil 1000Kbit burst 1724b cburst 1724b
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0) <- packet count is 0
rate 0bit 0pps backlog 0b 0p requeues 0
lended: 0 borrowed: 0 giants: 0
tokens: 13476 ctokens: 13476
class htb 1:20 root prio 0 rate 2000Kbit ceil 2000Kbit burst 1849b cburst 1849b
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
lended: 0 borrowed: 0 giants: 0
tokens: 7226 ctokens: 7226
class htb 1:30 root prio 0 rate 3000Kbit ceil 3000Kbit burst 1974b cburst 1974b
Sent 25804 bytes 214 pkt (dropped 0, overlimits 0 requeues 0)
rate 3992bit 4pps backlog 0b 0p requeues 0
lended: 214 borrowed: 0 giants: 0
tokens: 4851 ctokens: 4851
2) mount cls_cgroup and set root's net_cls.classid
# mount -t cgroup -onet_cls none /cgroup/net/
# cd /cgroup/net/
# ls
net_cls.classid notify_on_release release_agent tasks
# echo 0x10010 > net_cls.classid # I want to classify 1:10
# cat net_cls.classid
65552 # 65552 = 65536 + 16 = 0x10010
3) using network
4) check the packet which is classified to the class which is set to net_cls.classid.
# /home/usui/src/iproute2-2.6.29-1/tc/tc -s class show dev bond0
class htb 1:10 root prio 0 rate 1000Kbit ceil 1000Kbit burst 1724b cburst 1724b
Sent 11238 bytes 95 pkt (dropped 0, overlimits 0 requeues 0) <- increase packet count
rate 2944bit 3pps backlog 0b 0p requeues 0
lended: 95 borrowed: 0 giants: 0
tokens: 12601 ctokens: 12601
class htb 1:20 root prio 0 rate 2000Kbit ceil 2000Kbit burst 1849b cburst 1849b
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
lended: 0 borrowed: 0 giants: 0
tokens: 7226 ctokens: 7226
class htb 1:30 root prio 0 rate 3000Kbit ceil 3000Kbit burst 1974b cburst 1974b
Sent 46092 bytes 342 pkt (dropped 0, overlimits 0 requeues 0)
rate 320bit 0pps backlog 0b 0p requeues 0
lended: 342 borrowed: 0 giants: 0
tokens: 5017 ctokens: 5017
--
Minoru Usui <usui@mxm.nes.nec.co.jp>
^ permalink raw reply [flat|nested] 12+ messages in thread
end of thread, other threads:[~2009-06-22 18:05 UTC | newest]
Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-06-15 0:56 [HELP] for using class group classifier network controller 김재열
2009-06-15 0:56 ` 김재열
2009-06-15 2:36 ` Minoru Usui
2009-06-15 10:30 ` Thomas Graf
[not found] ` <20090615103020.GA8894-PfSpb0PWhxZc2C7mugBRk2EX/6BAtgUQ@public.gmane.org>
2009-06-22 18:05 ` [Devel] " Paul Menage
2009-06-22 18:05 ` Paul Menage
[not found] ` <20090615113609.db5c0613.usui-vEbCzmPPqHe45+QrQBaojngSJqDPrsil@public.gmane.org>
2009-06-15 10:30 ` Thomas Graf
2009-06-15 2:36 ` Minoru Usui
[not found] <9f8101c9ed80$b730b830$8310fe81@etri.info>
[not found] ` <9f8101c9ed80$b730b830$8310fe81-tRM8rG9f+DVhl2p70BpVqQ@public.gmane.org>
2009-06-15 8:18 ` Minoru Usui
2009-06-15 8:18 ` Minoru Usui
[not found] <d89801c9ef32$cdfcf480$8310fe81@etri.info>
[not found] ` <d89801c9ef32$cdfcf480$8310fe81-tRM8rG9f+DVhl2p70BpVqQ@public.gmane.org>
2009-06-18 4:13 ` Minoru Usui
2009-06-18 4:13 ` Minoru Usui
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.