xiaomi-sm8450/android_kernel_xiaomi_sm8450
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Linux-2.6.12-rc2

Browse Source 
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!
This commit is contained in:
Linus Torvalds
2005-04-16 15:20:36 -07:00
commit 1da177e4c3
17291 changed files with 6718755 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

508
net/sched/Kconfig Normal file
View File

@@ -0,0 +1,508 @@
#
# Traffic control configuration.
#
choice
prompt "Packet scheduler clock source"
depends on NET_SCHED
default NET_SCH_CLK_JIFFIES
help
Packet schedulers need a monotonic clock that increments at a static
rate. The kernel provides several suitable interfaces, each with
different properties:
- high resolution (us or better)
- fast to read (minimal locking, no i/o access)
- synchronized on all processors
- handles cpu clock frequency changes
but nothing provides all of the above.
config NET_SCH_CLK_JIFFIES
bool "Timer interrupt"
help
Say Y here if you want to use the timer interrupt (jiffies) as clock
source. This clock source is fast, synchronized on all processors and
handles cpu clock frequency changes, but its resolution is too low
for accurate shaping except at very low speed.
config NET_SCH_CLK_GETTIMEOFDAY
bool "gettimeofday"
help
Say Y here if you want to use gettimeofday as clock source. This clock
source has high resolution, is synchronized on all processors and
handles cpu clock frequency changes, but it is slow.
Choose this if you need a high resolution clock source but can't use
the CPU's cycle counter.
config NET_SCH_CLK_CPU
bool "CPU cycle counter"
depends on X86_TSC || X86_64 || ALPHA || SPARC64 || PPC64 || IA64
help
Say Y here if you want to use the CPU's cycle counter as clock source.
This is a cheap and high resolution clock source, but on some
architectures it is not synchronized on all processors and doesn't
handle cpu clock frequency changes.
The useable cycle counters are:
x86/x86_64 - Timestamp Counter
alpha - Cycle Counter
sparc64 - %ticks register
ppc64 - Time base
ia64 - Interval Time Counter
Choose this if your CPU's cycle counter is working properly.
endchoice
config NET_SCH_CBQ
tristate "CBQ packet scheduler"
depends on NET_SCHED
---help---
Say Y here if you want to use the Class-Based Queueing (CBQ) packet
scheduling algorithm for some of your network devices. This
algorithm classifies the waiting packets into a tree-like hierarchy
of classes; the leaves of this tree are in turn scheduled by
separate algorithms (called "disciplines" in this context).
See the top of <file:net/sched/sch_cbq.c> for references about the
CBQ algorithm.
CBQ is a commonly used scheduler, so if you're unsure, you should
say Y here. Then say Y to all the queueing algorithms below that you
want to use as CBQ disciplines. Then say Y to "Packet classifier
API" and say Y to all the classifiers you want to use; a classifier
is a routine that allows you to sort your outgoing traffic into
classes based on a certain criterion.
To compile this code as a module, choose M here: the
module will be called sch_cbq.
config NET_SCH_HTB
tristate "HTB packet scheduler"
depends on NET_SCHED
---help---
Say Y here if you want to use the Hierarchical Token Buckets (HTB)
packet scheduling algorithm for some of your network devices. See
<http://luxik.cdi.cz/~devik/qos/htb/> for complete manual and
in-depth articles.
HTB is very similar to the CBQ regarding its goals however is has
different properties and different algorithm.
To compile this code as a module, choose M here: the
module will be called sch_htb.
config NET_SCH_HFSC
tristate "HFSC packet scheduler"
depends on NET_SCHED
---help---
Say Y here if you want to use the Hierarchical Fair Service Curve
(HFSC) packet scheduling algorithm for some of your network devices.
To compile this code as a module, choose M here: the
module will be called sch_hfsc.
#tristate ' H-PFQ packet scheduler' CONFIG_NET_SCH_HPFQ
config NET_SCH_ATM
tristate "ATM pseudo-scheduler"
depends on NET_SCHED && ATM
---help---
Say Y here if you want to use the ATM pseudo-scheduler. This
provides a framework for invoking classifiers (aka "filters"), which
in turn select classes of this queuing discipline. Each class maps
the flow(s) it is handling to a given virtual circuit (see the top of
<file:net/sched/sch_atm.c>).
To compile this code as a module, choose M here: the
module will be called sch_atm.
config NET_SCH_PRIO
tristate "The simplest PRIO pseudoscheduler"
depends on NET_SCHED
help
Say Y here if you want to use an n-band priority queue packet
"scheduler" for some of your network devices or as a leaf discipline
for the CBQ scheduling algorithm. If unsure, say Y.
To compile this code as a module, choose M here: the
module will be called sch_prio.
config NET_SCH_RED
tristate "RED queue"
depends on NET_SCHED
help
Say Y here if you want to use the Random Early Detection (RED)
packet scheduling algorithm for some of your network devices (see
the top of <file:net/sched/sch_red.c> for details and references
about the algorithm).
To compile this code as a module, choose M here: the
module will be called sch_red.
config NET_SCH_SFQ
tristate "SFQ queue"
depends on NET_SCHED
---help---
Say Y here if you want to use the Stochastic Fairness Queueing (SFQ)
packet scheduling algorithm for some of your network devices or as a
leaf discipline for the CBQ scheduling algorithm (see the top of
<file:net/sched/sch_sfq.c> for details and references about the SFQ
algorithm).
To compile this code as a module, choose M here: the
module will be called sch_sfq.
config NET_SCH_TEQL
tristate "TEQL queue"
depends on NET_SCHED
---help---
Say Y here if you want to use the True Link Equalizer (TLE) packet
scheduling algorithm for some of your network devices or as a leaf
discipline for the CBQ scheduling algorithm. This queueing
discipline allows the combination of several physical devices into
one virtual device. (see the top of <file:net/sched/sch_teql.c> for
details).
To compile this code as a module, choose M here: the
module will be called sch_teql.
config NET_SCH_TBF
tristate "TBF queue"
depends on NET_SCHED
help
Say Y here if you want to use the Simple Token Bucket Filter (TBF)
packet scheduling algorithm for some of your network devices or as a
leaf discipline for the CBQ scheduling algorithm (see the top of
<file:net/sched/sch_tbf.c> for a description of the TBF algorithm).
To compile this code as a module, choose M here: the
module will be called sch_tbf.
config NET_SCH_GRED
tristate "GRED queue"
depends on NET_SCHED
help
Say Y here if you want to use the Generic Random Early Detection
(RED) packet scheduling algorithm for some of your network devices
(see the top of <file:net/sched/sch_red.c> for details and
references about the algorithm).
To compile this code as a module, choose M here: the
module will be called sch_gred.
config NET_SCH_DSMARK
tristate "Diffserv field marker"
depends on NET_SCHED
help
Say Y if you want to schedule packets according to the
Differentiated Services architecture proposed in RFC 2475.
Technical information on this method, with pointers to associated
RFCs, is available at <http://www.gta.ufrj.br/diffserv/>.
To compile this code as a module, choose M here: the
module will be called sch_dsmark.
config NET_SCH_NETEM
tristate "Network emulator"
depends on NET_SCHED
help
Say Y if you want to emulate network delay, loss, and packet
re-ordering. This is often useful to simulate networks when
testing applications or protocols.
To compile this driver as a module, choose M here: the module
will be called sch_netem.
If unsure, say N.
config NET_SCH_INGRESS
tristate "Ingress Qdisc"
depends on NET_SCHED
help
If you say Y here, you will be able to police incoming bandwidth
and drop packets when this bandwidth exceeds your desired rate.
If unsure, say Y.
To compile this code as a module, choose M here: the
module will be called sch_ingress.
config NET_QOS
bool "QoS support"
depends on NET_SCHED
---help---
Say Y here if you want to include Quality Of Service scheduling
features, which means that you will be able to request certain
rate-of-flow limits for your network devices.
This Quality of Service (QoS) support will enable you to use
Differentiated Services (diffserv) and Resource Reservation Protocol
(RSVP) on your Linux router if you also say Y to "Packet classifier
API" and to some classifiers below. Documentation and software is at
<http://diffserv.sourceforge.net/>.
Note that the answer to this question won't directly affect the
kernel: saying N will just cause the configurator to skip all
the questions about QoS support.
config NET_ESTIMATOR
bool "Rate estimator"
depends on NET_QOS
help
In order for Quality of Service scheduling to work, the current
rate-of-flow for a network device has to be estimated; if you say Y
here, the kernel will do just that.
config NET_CLS
bool "Packet classifier API"
depends on NET_SCHED
---help---
The CBQ scheduling algorithm requires that network packets which are
scheduled to be sent out over a network device be classified
according to some criterion. If you say Y here, you will get a
choice of several different packet classifiers with the following
questions.
This will enable you to use Differentiated Services (diffserv) and
Resource Reservation Protocol (RSVP) on your Linux router.
Documentation and software is at
<http://diffserv.sourceforge.net/>.
config NET_CLS_BASIC
tristate "Basic classifier"
depends on NET_CLS
---help---
Say Y here if you want to be able to classify packets using
only extended matches and actions.
To compile this code as a module, choose M here: the
module will be called cls_basic.
config NET_CLS_TCINDEX
tristate "TC index classifier"
depends on NET_CLS
help
If you say Y here, you will be able to classify outgoing packets
according to the tc_index field of the skb. You will want this
feature if you want to implement Differentiated Services using
sch_dsmark. If unsure, say Y.
To compile this code as a module, choose M here: the
module will be called cls_tcindex.
config NET_CLS_ROUTE4
tristate "Routing table based classifier"
depends on NET_CLS
select NET_CLS_ROUTE
help
If you say Y here, you will be able to classify outgoing packets
according to the route table entry they matched. If unsure, say Y.
To compile this code as a module, choose M here: the
module will be called cls_route.
config NET_CLS_ROUTE
bool
default n
config NET_CLS_FW
tristate "Firewall based classifier"
depends on NET_CLS
help
If you say Y here, you will be able to classify outgoing packets
according to firewall criteria you specified.
To compile this code as a module, choose M here: the
module will be called cls_fw.
config NET_CLS_U32
tristate "U32 classifier"
depends on NET_CLS
help
If you say Y here, you will be able to classify outgoing packets
according to their destination address. If unsure, say Y.
To compile this code as a module, choose M here: the
module will be called cls_u32.
config CLS_U32_PERF
bool "U32 classifier performance counters"
depends on NET_CLS_U32
help
gathers stats that could be used to tune u32 classifier performance.
Requires a new iproute2
You MUST NOT turn this on if you dont have an update iproute2.
config NET_CLS_IND
bool "classify input device (slows things u32/fw) "
depends on NET_CLS_U32 || NET_CLS_FW
help
This option will be killed eventually when a
metadata action appears because it slows things a little
Available only for u32 and fw classifiers.
Requires a new iproute2
You MUST NOT turn this on if you dont have an update iproute2.
config CLS_U32_MARK
bool "Use nfmark as a key in U32 classifier"
depends on NET_CLS_U32 && NETFILTER
help
This allows you to match mark in a u32 filter.
Example:
tc filter add dev eth0 protocol ip parent 1:0 prio 5 u32 \
match mark 0x0090 0xffff \
match ip dst 4.4.4.4 \
flowid 1:90
You must use a new iproute2 to use this feature.
config NET_CLS_RSVP
tristate "Special RSVP classifier"
depends on NET_CLS && NET_QOS
---help---
The Resource Reservation Protocol (RSVP) permits end systems to
request a minimum and maximum data flow rate for a connection; this
is important for real time data such as streaming sound or video.
Say Y here if you want to be able to classify outgoing packets based
on their RSVP requests.
To compile this code as a module, choose M here: the
module will be called cls_rsvp.
config NET_CLS_RSVP6
tristate "Special RSVP classifier for IPv6"
depends on NET_CLS && NET_QOS
---help---
The Resource Reservation Protocol (RSVP) permits end systems to
request a minimum and maximum data flow rate for a connection; this
is important for real time data such as streaming sound or video.
Say Y here if you want to be able to classify outgoing packets based
on their RSVP requests and you are using the new Internet Protocol
IPv6 as opposed to the older and more common IPv4.
To compile this code as a module, choose M here: the
module will be called cls_rsvp6.
config NET_EMATCH
bool "Extended Matches"
depends on NET_CLS
---help---
Say Y here if you want to use extended matches on top of classifiers
and select the extended matches below.
Extended matches are small classification helpers not worth writing
a separate classifier.
You must have a recent version of the iproute2 tools in order to use
extended matches.
config NET_EMATCH_STACK
int "Stack size"
depends on NET_EMATCH
default "32"
---help---
Size of the local stack variable used while evaluating the tree of
ematches. Limits the depth of the tree, i.e. the number of
encapsulated precedences. Every level requires 4 bytes of addtional
stack space.
config NET_EMATCH_CMP
tristate "Simple packet data comparison"
depends on NET_EMATCH
---help---
Say Y here if you want to be able to classify packets based on
simple packet data comparisons for 8, 16, and 32bit values.
To compile this code as a module, choose M here: the
module will be called em_cmp.
config NET_EMATCH_NBYTE
tristate "Multi byte comparison"
depends on NET_EMATCH
---help---
Say Y here if you want to be able to classify packets based on
multiple byte comparisons mainly useful for IPv6 address comparisons.
To compile this code as a module, choose M here: the
module will be called em_nbyte.
config NET_EMATCH_U32
tristate "U32 hashing key"
depends on NET_EMATCH
---help---
Say Y here if you want to be able to classify packets using
the famous u32 key in combination with logic relations.
To compile this code as a module, choose M here: the
module will be called em_u32.
config NET_EMATCH_META
tristate "Metadata"
depends on NET_EMATCH
---help---
Say Y here if you want to be ablt to classify packets based on
metadata such as load average, netfilter attributes, socket
attributes and routing decisions.
To compile this code as a module, choose M here: the
module will be called em_meta.
config NET_CLS_ACT
bool "Packet ACTION"
depends on EXPERIMENTAL && NET_CLS && NET_QOS
---help---
This option requires you have a new iproute2. It enables
tc extensions which can be used with tc classifiers.
You MUST NOT turn this on if you dont have an update iproute2.
config NET_ACT_POLICE
tristate "Policing Actions"
depends on NET_CLS_ACT
---help---
If you are using a newer iproute2 select this one, otherwise use one
below to select a policer.
You MUST NOT turn this on if you dont have an update iproute2.
config NET_ACT_GACT
tristate "generic Actions"
depends on NET_CLS_ACT
---help---
You must have new iproute2 to use this feature.
This adds simple filtering actions like drop, accept etc.
config GACT_PROB
bool "generic Actions probability"
depends on NET_ACT_GACT
---help---
Allows generic actions to be randomly or deterministically used.
config NET_ACT_MIRRED
tristate "Packet In/Egress redirecton/mirror Actions"
depends on NET_CLS_ACT
---help---
requires new iproute2
This allows packets to be mirrored or redirected to netdevices
config NET_ACT_IPT
tristate "iptables Actions"
depends on NET_CLS_ACT && NETFILTER && IP_NF_IPTABLES
---help---
requires new iproute2
This allows iptables targets to be used by tc filters
config NET_ACT_PEDIT
tristate "Generic Packet Editor Actions"
depends on NET_CLS_ACT
---help---
requires new iproute2
This allows for packets to be generically edited
config NET_CLS_POLICE
bool "Traffic policing (needed for in/egress)"
depends on NET_CLS && NET_QOS && NET_CLS_ACT!=y
help
Say Y to support traffic policing (bandwidth limits). Needed for
ingress and egress rate limiting.

41
net/sched/Makefile Normal file
View File

@@ -0,0 +1,41 @@
#
# Makefile for the Linux Traffic Control Unit.
#
obj-y := sch_generic.o
obj-$(CONFIG_NET_SCHED) += sch_api.o sch_fifo.o
obj-$(CONFIG_NET_CLS) += cls_api.o
obj-$(CONFIG_NET_CLS_ACT) += act_api.o
obj-$(CONFIG_NET_ACT_POLICE) += police.o
obj-$(CONFIG_NET_CLS_POLICE) += police.o
obj-$(CONFIG_NET_ACT_GACT) += gact.o
obj-$(CONFIG_NET_ACT_MIRRED) += mirred.o
obj-$(CONFIG_NET_ACT_IPT) += ipt.o
obj-$(CONFIG_NET_ACT_PEDIT) += pedit.o
obj-$(CONFIG_NET_SCH_CBQ) += sch_cbq.o
obj-$(CONFIG_NET_SCH_HTB) += sch_htb.o
obj-$(CONFIG_NET_SCH_HPFQ) += sch_hpfq.o
obj-$(CONFIG_NET_SCH_HFSC) += sch_hfsc.o
obj-$(CONFIG_NET_SCH_RED) += sch_red.o
obj-$(CONFIG_NET_SCH_GRED) += sch_gred.o
obj-$(CONFIG_NET_SCH_INGRESS) += sch_ingress.o
obj-$(CONFIG_NET_SCH_DSMARK) += sch_dsmark.o
obj-$(CONFIG_NET_SCH_SFQ) += sch_sfq.o
obj-$(CONFIG_NET_SCH_TBF) += sch_tbf.o
obj-$(CONFIG_NET_SCH_TEQL) += sch_teql.o
obj-$(CONFIG_NET_SCH_PRIO) += sch_prio.o
obj-$(CONFIG_NET_SCH_ATM) += sch_atm.o
obj-$(CONFIG_NET_SCH_NETEM) += sch_netem.o
obj-$(CONFIG_NET_CLS_U32) += cls_u32.o
obj-$(CONFIG_NET_CLS_ROUTE4) += cls_route.o
obj-$(CONFIG_NET_CLS_FW) += cls_fw.o
obj-$(CONFIG_NET_CLS_RSVP) += cls_rsvp.o
obj-$(CONFIG_NET_CLS_TCINDEX) += cls_tcindex.o
obj-$(CONFIG_NET_CLS_RSVP6) += cls_rsvp6.o
obj-$(CONFIG_NET_CLS_BASIC) += cls_basic.o
obj-$(CONFIG_NET_EMATCH) += ematch.o
obj-$(CONFIG_NET_EMATCH_CMP) += em_cmp.o
obj-$(CONFIG_NET_EMATCH_NBYTE) += em_nbyte.o
obj-$(CONFIG_NET_EMATCH_U32) += em_u32.o
obj-$(CONFIG_NET_EMATCH_META) += em_meta.o

894
net/sched/act_api.c Normal file
View File

@@ -0,0 +1,894 @@
/*
* net/sched/act_api.c Packet action API.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Author: Jamal Hadi Salim
*
*
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/config.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/init.h>
#include <linux/kmod.h>
#include <net/sock.h>
#include <net/sch_generic.h>
#include <net/act_api.h>
#if 1 /* control */
#define DPRINTK(format, args...) printk(KERN_DEBUG format, ##args)
#else
#define DPRINTK(format, args...)
#endif
#if 0 /* data */
#define D2PRINTK(format, args...) printk(KERN_DEBUG format, ##args)
#else
#define D2PRINTK(format, args...)
#endif
static struct tc_action_ops *act_base = NULL;
static DEFINE_RWLOCK(act_mod_lock);
int tcf_register_action(struct tc_action_ops *act)
{
struct tc_action_ops *a, **ap;
write_lock(&act_mod_lock);
for (ap = &act_base; (a = *ap) != NULL; ap = &a->next) {
if (act->type == a->type || (strcmp(act->kind, a->kind) == 0)) {
write_unlock(&act_mod_lock);
return -EEXIST;
}
}
act->next = NULL;
*ap = act;
write_unlock(&act_mod_lock);
return 0;
}
int tcf_unregister_action(struct tc_action_ops *act)
{
struct tc_action_ops *a, **ap;
int err = -ENOENT;
write_lock(&act_mod_lock);
for (ap = &act_base; (a = *ap) != NULL; ap = &a->next)
if (a == act)
break;
if (a) {
*ap = a->next;
a->next = NULL;
err = 0;
}
write_unlock(&act_mod_lock);
return err;
}
/* lookup by name */
static struct tc_action_ops *tc_lookup_action_n(char *kind)
{
struct tc_action_ops *a = NULL;
if (kind) {
read_lock(&act_mod_lock);
for (a = act_base; a; a = a->next) {
if (strcmp(kind, a->kind) == 0) {
if (!try_module_get(a->owner)) {
read_unlock(&act_mod_lock);
return NULL;
}
break;
}
}
read_unlock(&act_mod_lock);
}
return a;
}
/* lookup by rtattr */
static struct tc_action_ops *tc_lookup_action(struct rtattr *kind)
{
struct tc_action_ops *a = NULL;
if (kind) {
read_lock(&act_mod_lock);
for (a = act_base; a; a = a->next) {
if (rtattr_strcmp(kind, a->kind) == 0) {
if (!try_module_get(a->owner)) {
read_unlock(&act_mod_lock);
return NULL;
}
break;
}
}
read_unlock(&act_mod_lock);
}
return a;
}
#if 0
/* lookup by id */
static struct tc_action_ops *tc_lookup_action_id(u32 type)
{
struct tc_action_ops *a = NULL;
if (type) {
read_lock(&act_mod_lock);
for (a = act_base; a; a = a->next) {
if (a->type == type) {
if (!try_module_get(a->owner)) {
read_unlock(&act_mod_lock);
return NULL;
}
break;
}
}
read_unlock(&act_mod_lock);
}
return a;
}
#endif
int tcf_action_exec(struct sk_buff *skb, struct tc_action *act,
struct tcf_result *res)
{
struct tc_action *a;
int ret = -1;
if (skb->tc_verd & TC_NCLS) {
skb->tc_verd = CLR_TC_NCLS(skb->tc_verd);
D2PRINTK("(%p)tcf_action_exec: cleared TC_NCLS in %s out %s\n",
skb, skb->input_dev ? skb->input_dev->name : "xxx",
skb->dev->name);
ret = TC_ACT_OK;
goto exec_done;
}
while ((a = act) != NULL) {
repeat:
if (a->ops && a->ops->act) {
ret = a->ops->act(&skb, a);
if (TC_MUNGED & skb->tc_verd) {
/* copied already, allow trampling */
skb->tc_verd = SET_TC_OK2MUNGE(skb->tc_verd);
skb->tc_verd = CLR_TC_MUNGED(skb->tc_verd);
}
if (ret != TC_ACT_PIPE)
goto exec_done;
if (ret == TC_ACT_REPEAT)
goto repeat; /* we need a ttl - JHS */
}
act = a->next;
}
exec_done:
if (skb->tc_classid > 0) {
res->classid = skb->tc_classid;
res->class = 0;
skb->tc_classid = 0;
}
return ret;
}
void tcf_action_destroy(struct tc_action *act, int bind)
{
struct tc_action *a;
for (a = act; a; a = act) {
if (a->ops && a->ops->cleanup) {
DPRINTK("tcf_action_destroy destroying %p next %p\n",
a, a->next);
if (a->ops->cleanup(a, bind) == ACT_P_DELETED)
module_put(a->ops->owner);
act = act->next;
kfree(a);
} else { /*FIXME: Remove later - catch insertion bugs*/
printk("tcf_action_destroy: BUG? destroying NULL ops\n");
act = act->next;
kfree(a);
}
}
}
int
tcf_action_dump_old(struct sk_buff *skb, struct tc_action *a, int bind, int ref)
{
int err = -EINVAL;
if (a->ops == NULL || a->ops->dump == NULL)
return err;
return a->ops->dump(skb, a, bind, ref);
}
int
tcf_action_dump_1(struct sk_buff *skb, struct tc_action *a, int bind, int ref)
{
int err = -EINVAL;
unsigned char *b = skb->tail;
struct rtattr *r;
if (a->ops == NULL || a->ops->dump == NULL)
return err;
RTA_PUT(skb, TCA_KIND, IFNAMSIZ, a->ops->kind);
if (tcf_action_copy_stats(skb, a, 0))
goto rtattr_failure;
r = (struct rtattr*) skb->tail;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
if ((err = tcf_action_dump_old(skb, a, bind, ref)) > 0) {
r->rta_len = skb->tail - (u8*)r;
return err;
}
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
int
tcf_action_dump(struct sk_buff *skb, struct tc_action *act, int bind, int ref)
{
struct tc_action *a;
int err = -EINVAL;
unsigned char *b = skb->tail;
struct rtattr *r ;
while ((a = act) != NULL) {
r = (struct rtattr*) skb->tail;
act = a->next;
RTA_PUT(skb, a->order, 0, NULL);
err = tcf_action_dump_1(skb, a, bind, ref);
if (err < 0)
goto rtattr_failure;
r->rta_len = skb->tail - (u8*)r;
}
return 0;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -err;
}
struct tc_action *tcf_action_init_1(struct rtattr *rta, struct rtattr *est,
char *name, int ovr, int bind, int *err)
{
struct tc_action *a;
struct tc_action_ops *a_o;
char act_name[IFNAMSIZ];
struct rtattr *tb[TCA_ACT_MAX+1];
struct rtattr *kind;
*err = -EINVAL;
if (name == NULL) {
if (rtattr_parse_nested(tb, TCA_ACT_MAX, rta) < 0)
goto err_out;
kind = tb[TCA_ACT_KIND-1];
if (kind == NULL)
goto err_out;
if (rtattr_strlcpy(act_name, kind, IFNAMSIZ) >= IFNAMSIZ)
goto err_out;
} else {
if (strlcpy(act_name, name, IFNAMSIZ) >= IFNAMSIZ)
goto err_out;
}
a_o = tc_lookup_action_n(act_name);
if (a_o == NULL) {
#ifdef CONFIG_KMOD
rtnl_unlock();
request_module(act_name);
rtnl_lock();
a_o = tc_lookup_action_n(act_name);
/* We dropped the RTNL semaphore in order to
* perform the module load. So, even if we
* succeeded in loading the module we have to
* tell the caller to replay the request. We
* indicate this using -EAGAIN.
*/
if (a_o != NULL) {
*err = -EAGAIN;
goto err_mod;
}
#endif
goto err_out;
}
*err = -ENOMEM;
a = kmalloc(sizeof(*a), GFP_KERNEL);
if (a == NULL)
goto err_mod;
memset(a, 0, sizeof(*a));
/* backward compatibility for policer */
if (name == NULL)
*err = a_o->init(tb[TCA_ACT_OPTIONS-1], est, a, ovr, bind);
else
*err = a_o->init(rta, est, a, ovr, bind);
if (*err < 0)
goto err_free;
/* module count goes up only when brand new policy is created
if it exists and is only bound to in a_o->init() then
ACT_P_CREATED is not returned (a zero is).
*/
if (*err != ACT_P_CREATED)
module_put(a_o->owner);
a->ops = a_o;
DPRINTK("tcf_action_init_1: successfull %s\n", act_name);
*err = 0;
return a;
err_free:
kfree(a);
err_mod:
module_put(a_o->owner);
err_out:
return NULL;
}
struct tc_action *tcf_action_init(struct rtattr *rta, struct rtattr *est,
char *name, int ovr, int bind, int *err)
{
struct rtattr *tb[TCA_ACT_MAX_PRIO+1];
struct tc_action *head = NULL, *act, *act_prev = NULL;
int i;
if (rtattr_parse_nested(tb, TCA_ACT_MAX_PRIO, rta) < 0) {
*err = -EINVAL;
return head;
}
for (i=0; i < TCA_ACT_MAX_PRIO && tb[i]; i++) {
act = tcf_action_init_1(tb[i], est, name, ovr, bind, err);
if (act == NULL)
goto err;
act->order = i+1;
if (head == NULL)
head = act;
else
act_prev->next = act;
act_prev = act;
}
return head;
err:
if (head != NULL)
tcf_action_destroy(head, bind);
return NULL;
}
int tcf_action_copy_stats(struct sk_buff *skb, struct tc_action *a,
int compat_mode)
{
int err = 0;
struct gnet_dump d;
struct tcf_act_hdr *h = a->priv;
if (h == NULL)
goto errout;
/* compat_mode being true specifies a call that is supposed
* to add additional backward compatiblity statistic TLVs.
*/
if (compat_mode) {
if (a->type == TCA_OLD_COMPAT)
err = gnet_stats_start_copy_compat(skb, 0,
TCA_STATS, TCA_XSTATS, h->stats_lock, &d);
else
return 0;
} else
err = gnet_stats_start_copy(skb, TCA_ACT_STATS,
h->stats_lock, &d);
if (err < 0)
goto errout;
if (a->ops != NULL && a->ops->get_stats != NULL)
if (a->ops->get_stats(skb, a) < 0)
goto errout;
if (gnet_stats_copy_basic(&d, &h->bstats) < 0 ||
#ifdef CONFIG_NET_ESTIMATOR
gnet_stats_copy_rate_est(&d, &h->rate_est) < 0 ||
#endif
gnet_stats_copy_queue(&d, &h->qstats) < 0)
goto errout;
if (gnet_stats_finish_copy(&d) < 0)
goto errout;
return 0;
errout:
return -1;
}
static int
tca_get_fill(struct sk_buff *skb, struct tc_action *a, u32 pid, u32 seq,
unsigned flags, int event, int bind, int ref)
{
struct tcamsg *t;
struct nlmsghdr *nlh;
unsigned char *b = skb->tail;
struct rtattr *x;
nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*t));
nlh->nlmsg_flags = flags;
t = NLMSG_DATA(nlh);
t->tca_family = AF_UNSPEC;
x = (struct rtattr*) skb->tail;
RTA_PUT(skb, TCA_ACT_TAB, 0, NULL);
if (tcf_action_dump(skb, a, bind, ref) < 0)
goto rtattr_failure;
x->rta_len = skb->tail - (u8*)x;
nlh->nlmsg_len = skb->tail - b;
return skb->len;
rtattr_failure:
nlmsg_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int
act_get_notify(u32 pid, struct nlmsghdr *n, struct tc_action *a, int event)
{
struct sk_buff *skb;
int err = 0;
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
if (!skb)
return -ENOBUFS;
if (tca_get_fill(skb, a, pid, n->nlmsg_seq, 0, event, 0, 0) <= 0) {
kfree_skb(skb);
return -EINVAL;
}
err = netlink_unicast(rtnl, skb, pid, MSG_DONTWAIT);
if (err > 0)
err = 0;
return err;
}
static struct tc_action *
tcf_action_get_1(struct rtattr *rta, struct nlmsghdr *n, u32 pid, int *err)
{
struct rtattr *tb[TCA_ACT_MAX+1];
struct tc_action *a;
int index;
*err = -EINVAL;
if (rtattr_parse_nested(tb, TCA_ACT_MAX, rta) < 0)
return NULL;
if (tb[TCA_ACT_INDEX - 1] == NULL ||
RTA_PAYLOAD(tb[TCA_ACT_INDEX - 1]) < sizeof(index))
return NULL;
index = *(int *)RTA_DATA(tb[TCA_ACT_INDEX - 1]);
*err = -ENOMEM;
a = kmalloc(sizeof(struct tc_action), GFP_KERNEL);
if (a == NULL)
return NULL;
memset(a, 0, sizeof(struct tc_action));
*err = -EINVAL;
a->ops = tc_lookup_action(tb[TCA_ACT_KIND - 1]);
if (a->ops == NULL)
goto err_free;
if (a->ops->lookup == NULL)
goto err_mod;
*err = -ENOENT;
if (a->ops->lookup(a, index) == 0)
goto err_mod;
module_put(a->ops->owner);
*err = 0;
return a;
err_mod:
module_put(a->ops->owner);
err_free:
kfree(a);
return NULL;
}
static void cleanup_a(struct tc_action *act)
{
struct tc_action *a;
for (a = act; a; a = act) {
act = a->next;
kfree(a);
}
}
static struct tc_action *create_a(int i)
{
struct tc_action *act;
act = kmalloc(sizeof(*act), GFP_KERNEL);
if (act == NULL) {
printk("create_a: failed to alloc!\n");
return NULL;
}
memset(act, 0, sizeof(*act));
act->order = i;
return act;
}
static int tca_action_flush(struct rtattr *rta, struct nlmsghdr *n, u32 pid)
{
struct sk_buff *skb;
unsigned char *b;
struct nlmsghdr *nlh;
struct tcamsg *t;
struct netlink_callback dcb;
struct rtattr *x;
struct rtattr *tb[TCA_ACT_MAX+1];
struct rtattr *kind;
struct tc_action *a = create_a(0);
int err = -EINVAL;
if (a == NULL) {
printk("tca_action_flush: couldnt create tc_action\n");
return err;
}
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
if (!skb) {
printk("tca_action_flush: failed skb alloc\n");
kfree(a);
return -ENOBUFS;
}
b = (unsigned char *)skb->tail;
if (rtattr_parse_nested(tb, TCA_ACT_MAX, rta) < 0)
goto err_out;
kind = tb[TCA_ACT_KIND-1];
a->ops = tc_lookup_action(kind);
if (a->ops == NULL)
goto err_out;
nlh = NLMSG_PUT(skb, pid, n->nlmsg_seq, RTM_DELACTION, sizeof(*t));
t = NLMSG_DATA(nlh);
t->tca_family = AF_UNSPEC;
x = (struct rtattr *) skb->tail;
RTA_PUT(skb, TCA_ACT_TAB, 0, NULL);
err = a->ops->walk(skb, &dcb, RTM_DELACTION, a);
if (err < 0)
goto rtattr_failure;
x->rta_len = skb->tail - (u8 *) x;
nlh->nlmsg_len = skb->tail - b;
nlh->nlmsg_flags |= NLM_F_ROOT;
module_put(a->ops->owner);
kfree(a);
err = rtnetlink_send(skb, pid, RTMGRP_TC, n->nlmsg_flags&NLM_F_ECHO);
if (err > 0)
return 0;
return err;
rtattr_failure:
module_put(a->ops->owner);
nlmsg_failure:
err_out:
kfree_skb(skb);
kfree(a);
return err;
}
static int
tca_action_gd(struct rtattr *rta, struct nlmsghdr *n, u32 pid, int event)
{
int i, ret = 0;
struct rtattr *tb[TCA_ACT_MAX_PRIO+1];
struct tc_action *head = NULL, *act, *act_prev = NULL;
if (rtattr_parse_nested(tb, TCA_ACT_MAX_PRIO, rta) < 0)
return -EINVAL;
if (event == RTM_DELACTION && n->nlmsg_flags&NLM_F_ROOT) {
if (tb[0] != NULL && tb[1] == NULL)
return tca_action_flush(tb[0], n, pid);
}
for (i=0; i < TCA_ACT_MAX_PRIO && tb[i]; i++) {
act = tcf_action_get_1(tb[i], n, pid, &ret);
if (act == NULL)
goto err;
act->order = i+1;
if (head == NULL)
head = act;
else
act_prev->next = act;
act_prev = act;
}
if (event == RTM_GETACTION)
ret = act_get_notify(pid, n, head, event);
else { /* delete */
struct sk_buff *skb;
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
if (!skb) {
ret = -ENOBUFS;
goto err;
}
if (tca_get_fill(skb, head, pid, n->nlmsg_seq, 0, event,
0, 1) <= 0) {
kfree_skb(skb);
ret = -EINVAL;
goto err;
}
/* now do the delete */
tcf_action_destroy(head, 0);
ret = rtnetlink_send(skb, pid, RTMGRP_TC,
n->nlmsg_flags&NLM_F_ECHO);
if (ret > 0)
return 0;
return ret;
}
err:
cleanup_a(head);
return ret;
}
static int tcf_add_notify(struct tc_action *a, u32 pid, u32 seq, int event,
unsigned flags)
{
struct tcamsg *t;
struct nlmsghdr *nlh;
struct sk_buff *skb;
struct rtattr *x;
unsigned char *b;
int err = 0;
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
if (!skb)
return -ENOBUFS;
b = (unsigned char *)skb->tail;
nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*t));
nlh->nlmsg_flags = flags;
t = NLMSG_DATA(nlh);
t->tca_family = AF_UNSPEC;
x = (struct rtattr*) skb->tail;
RTA_PUT(skb, TCA_ACT_TAB, 0, NULL);
if (tcf_action_dump(skb, a, 0, 0) < 0)
goto rtattr_failure;
x->rta_len = skb->tail - (u8*)x;
nlh->nlmsg_len = skb->tail - b;
NETLINK_CB(skb).dst_groups = RTMGRP_TC;
err = rtnetlink_send(skb, pid, RTMGRP_TC, flags&NLM_F_ECHO);
if (err > 0)
err = 0;
return err;
rtattr_failure:
nlmsg_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int
tcf_action_add(struct rtattr *rta, struct nlmsghdr *n, u32 pid, int ovr)
{
int ret = 0;
struct tc_action *act;
struct tc_action *a;
u32 seq = n->nlmsg_seq;
act = tcf_action_init(rta, NULL, NULL, ovr, 0, &ret);
if (act == NULL)
goto done;
/* dump then free all the actions after update; inserted policy
* stays intact
* */
ret = tcf_add_notify(act, pid, seq, RTM_NEWACTION, n->nlmsg_flags);
for (a = act; a; a = act) {
act = a->next;
kfree(a);
}
done:
return ret;
}
static int tc_ctl_action(struct sk_buff *skb, struct nlmsghdr *n, void *arg)
{
struct rtattr **tca = arg;
u32 pid = skb ? NETLINK_CB(skb).pid : 0;
int ret = 0, ovr = 0;
if (tca[TCA_ACT_TAB-1] == NULL) {
printk("tc_ctl_action: received NO action attribs\n");
return -EINVAL;
}
/* n->nlmsg_flags&NLM_F_CREATE
* */
switch (n->nlmsg_type) {
case RTM_NEWACTION:
/* we are going to assume all other flags
* imply create only if it doesnt exist
* Note that CREATE | EXCL implies that
* but since we want avoid ambiguity (eg when flags
* is zero) then just set this
*/
if (n->nlmsg_flags&NLM_F_REPLACE)
ovr = 1;
replay:
ret = tcf_action_add(tca[TCA_ACT_TAB-1], n, pid, ovr);
if (ret == -EAGAIN)
goto replay;
break;
case RTM_DELACTION:
ret = tca_action_gd(tca[TCA_ACT_TAB-1], n, pid, RTM_DELACTION);
break;
case RTM_GETACTION:
ret = tca_action_gd(tca[TCA_ACT_TAB-1], n, pid, RTM_GETACTION);
break;
default:
BUG();
}
return ret;
}
static char *
find_dump_kind(struct nlmsghdr *n)
{
struct rtattr *tb1, *tb2[TCA_ACT_MAX+1];
struct rtattr *tb[TCA_ACT_MAX_PRIO + 1];
struct rtattr *rta[TCAA_MAX + 1];
struct rtattr *kind;
int min_len = NLMSG_LENGTH(sizeof(struct tcamsg));
int attrlen = n->nlmsg_len - NLMSG_ALIGN(min_len);
struct rtattr *attr = (void *) n + NLMSG_ALIGN(min_len);
if (rtattr_parse(rta, TCAA_MAX, attr, attrlen) < 0)
return NULL;
tb1 = rta[TCA_ACT_TAB - 1];
if (tb1 == NULL)
return NULL;
if (rtattr_parse(tb, TCA_ACT_MAX_PRIO, RTA_DATA(tb1),
NLMSG_ALIGN(RTA_PAYLOAD(tb1))) < 0)
return NULL;
if (tb[0] == NULL)
return NULL;
if (rtattr_parse(tb2, TCA_ACT_MAX, RTA_DATA(tb[0]),
RTA_PAYLOAD(tb[0])) < 0)
return NULL;
kind = tb2[TCA_ACT_KIND-1];
return (char *) RTA_DATA(kind);
}
static int
tc_dump_action(struct sk_buff *skb, struct netlink_callback *cb)
{
struct nlmsghdr *nlh;
unsigned char *b = skb->tail;
struct rtattr *x;
struct tc_action_ops *a_o;
struct tc_action a;
int ret = 0;
struct tcamsg *t = (struct tcamsg *) NLMSG_DATA(cb->nlh);
char *kind = find_dump_kind(cb->nlh);
if (kind == NULL) {
printk("tc_dump_action: action bad kind\n");
return 0;
}
a_o = tc_lookup_action_n(kind);
if (a_o == NULL) {
printk("failed to find %s\n", kind);
return 0;
}
memset(&a, 0, sizeof(struct tc_action));
a.ops = a_o;
if (a_o->walk == NULL) {
printk("tc_dump_action: %s !capable of dumping table\n", kind);
goto rtattr_failure;
}
nlh = NLMSG_PUT(skb, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq,
cb->nlh->nlmsg_type, sizeof(*t));
t = NLMSG_DATA(nlh);
t->tca_family = AF_UNSPEC;
x = (struct rtattr *) skb->tail;
RTA_PUT(skb, TCA_ACT_TAB, 0, NULL);
ret = a_o->walk(skb, cb, RTM_GETACTION, &a);
if (ret < 0)
goto rtattr_failure;
if (ret > 0) {
x->rta_len = skb->tail - (u8 *) x;
ret = skb->len;
} else
skb_trim(skb, (u8*)x - skb->data);
nlh->nlmsg_len = skb->tail - b;
if (NETLINK_CB(cb->skb).pid && ret)
nlh->nlmsg_flags |= NLM_F_MULTI;
module_put(a_o->owner);
return skb->len;
rtattr_failure:
nlmsg_failure:
module_put(a_o->owner);
skb_trim(skb, b - skb->data);
return skb->len;
}
static int __init tc_action_init(void)
{
struct rtnetlink_link *link_p = rtnetlink_links[PF_UNSPEC];
if (link_p) {
link_p[RTM_NEWACTION-RTM_BASE].doit = tc_ctl_action;
link_p[RTM_DELACTION-RTM_BASE].doit = tc_ctl_action;
link_p[RTM_GETACTION-RTM_BASE].doit = tc_ctl_action;
link_p[RTM_GETACTION-RTM_BASE].dumpit = tc_dump_action;
}
printk("TC classifier action (bugs to netdev@oss.sgi.com cc "
"hadi@cyberus.ca)\n");
return 0;
}
subsys_initcall(tc_action_init);
EXPORT_SYMBOL(tcf_register_action);
EXPORT_SYMBOL(tcf_unregister_action);
EXPORT_SYMBOL(tcf_action_exec);
EXPORT_SYMBOL(tcf_action_dump_1);

642
net/sched/cls_api.c Normal file
View File

@@ -0,0 +1,642 @@
/*
* net/sched/cls_api.c Packet classifier API.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
* Changes:
*
* Eduardo J. Blanco <ejbs@netlabs.com.uy> :990222: kmod support
*
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/init.h>
#include <linux/kmod.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
#include <net/pkt_cls.h>
#if 0 /* control */
#define DPRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define DPRINTK(format,args...)
#endif
/* The list of all installed classifier types */
static struct tcf_proto_ops *tcf_proto_base;
/* Protects list of registered TC modules. It is pure SMP lock. */
static DEFINE_RWLOCK(cls_mod_lock);
/* Find classifier type by string name */
static struct tcf_proto_ops * tcf_proto_lookup_ops(struct rtattr *kind)
{
struct tcf_proto_ops *t = NULL;
if (kind) {
read_lock(&cls_mod_lock);
for (t = tcf_proto_base; t; t = t->next) {
if (rtattr_strcmp(kind, t->kind) == 0) {
if (!try_module_get(t->owner))
t = NULL;
break;
}
}
read_unlock(&cls_mod_lock);
}
return t;
}
/* Register(unregister) new classifier type */
int register_tcf_proto_ops(struct tcf_proto_ops *ops)
{
struct tcf_proto_ops *t, **tp;
int rc = -EEXIST;
write_lock(&cls_mod_lock);
for (tp = &tcf_proto_base; (t = *tp) != NULL; tp = &t->next)
if (!strcmp(ops->kind, t->kind))
goto out;
ops->next = NULL;
*tp = ops;
rc = 0;
out:
write_unlock(&cls_mod_lock);
return rc;
}
int unregister_tcf_proto_ops(struct tcf_proto_ops *ops)
{
struct tcf_proto_ops *t, **tp;
int rc = -ENOENT;
write_lock(&cls_mod_lock);
for (tp = &tcf_proto_base; (t=*tp) != NULL; tp = &t->next)
if (t == ops)
break;
if (!t)
goto out;
*tp = t->next;
rc = 0;
out:
write_unlock(&cls_mod_lock);
return rc;
}
static int tfilter_notify(struct sk_buff *oskb, struct nlmsghdr *n,
struct tcf_proto *tp, unsigned long fh, int event);
/* Select new prio value from the range, managed by kernel. */
static __inline__ u32 tcf_auto_prio(struct tcf_proto *tp)
{
u32 first = TC_H_MAKE(0xC0000000U,0U);
if (tp)
first = tp->prio-1;
return first;
}
/* Add/change/delete/get a filter node */
static int tc_ctl_tfilter(struct sk_buff *skb, struct nlmsghdr *n, void *arg)
{
struct rtattr **tca;
struct tcmsg *t;
u32 protocol;
u32 prio;
u32 nprio;
u32 parent;
struct net_device *dev;
struct Qdisc *q;
struct tcf_proto **back, **chain;
struct tcf_proto *tp;
struct tcf_proto_ops *tp_ops;
struct Qdisc_class_ops *cops;
unsigned long cl;
unsigned long fh;
int err;
replay:
tca = arg;
t = NLMSG_DATA(n);
protocol = TC_H_MIN(t->tcm_info);
prio = TC_H_MAJ(t->tcm_info);
nprio = prio;
parent = t->tcm_parent;
cl = 0;
if (prio == 0) {
/* If no priority is given, user wants we allocated it. */
if (n->nlmsg_type != RTM_NEWTFILTER || !(n->nlmsg_flags&NLM_F_CREATE))
return -ENOENT;
prio = TC_H_MAKE(0x80000000U,0U);
}
/* Find head of filter chain. */
/* Find link */
if ((dev = __dev_get_by_index(t->tcm_ifindex)) == NULL)
return -ENODEV;
/* Find qdisc */
if (!parent) {
q = dev->qdisc_sleeping;
parent = q->handle;
} else if ((q = qdisc_lookup(dev, TC_H_MAJ(t->tcm_parent))) == NULL)
return -EINVAL;
/* Is it classful? */
if ((cops = q->ops->cl_ops) == NULL)
return -EINVAL;
/* Do we search for filter, attached to class? */
if (TC_H_MIN(parent)) {
cl = cops->get(q, parent);
if (cl == 0)
return -ENOENT;
}
/* And the last stroke */
chain = cops->tcf_chain(q, cl);
err = -EINVAL;
if (chain == NULL)
goto errout;
/* Check the chain for existence of proto-tcf with this priority */
for (back = chain; (tp=*back) != NULL; back = &tp->next) {
if (tp->prio >= prio) {
if (tp->prio == prio) {
if (!nprio || (tp->protocol != protocol && protocol))
goto errout;
} else
tp = NULL;
break;
}
}
if (tp == NULL) {
/* Proto-tcf does not exist, create new one */
if (tca[TCA_KIND-1] == NULL || !protocol)
goto errout;
err = -ENOENT;
if (n->nlmsg_type != RTM_NEWTFILTER || !(n->nlmsg_flags&NLM_F_CREATE))
goto errout;
/* Create new proto tcf */
err = -ENOBUFS;
if ((tp = kmalloc(sizeof(*tp), GFP_KERNEL)) == NULL)
goto errout;
err = -EINVAL;
tp_ops = tcf_proto_lookup_ops(tca[TCA_KIND-1]);
if (tp_ops == NULL) {
#ifdef CONFIG_KMOD
struct rtattr *kind = tca[TCA_KIND-1];
char name[IFNAMSIZ];
if (kind != NULL &&
rtattr_strlcpy(name, kind, IFNAMSIZ) < IFNAMSIZ) {
rtnl_unlock();
request_module("cls_%s", name);
rtnl_lock();
tp_ops = tcf_proto_lookup_ops(kind);
/* We dropped the RTNL semaphore in order to
* perform the module load. So, even if we
* succeeded in loading the module we have to
* replay the request. We indicate this using
* -EAGAIN.
*/
if (tp_ops != NULL) {
module_put(tp_ops->owner);
err = -EAGAIN;
}
}
#endif
kfree(tp);
goto errout;
}
memset(tp, 0, sizeof(*tp));
tp->ops = tp_ops;
tp->protocol = protocol;
tp->prio = nprio ? : tcf_auto_prio(*back);
tp->q = q;
tp->classify = tp_ops->classify;
tp->classid = parent;
if ((err = tp_ops->init(tp)) != 0) {
module_put(tp_ops->owner);
kfree(tp);
goto errout;
}
qdisc_lock_tree(dev);
tp->next = *back;
*back = tp;
qdisc_unlock_tree(dev);
} else if (tca[TCA_KIND-1] && rtattr_strcmp(tca[TCA_KIND-1], tp->ops->kind))
goto errout;
fh = tp->ops->get(tp, t->tcm_handle);
if (fh == 0) {
if (n->nlmsg_type == RTM_DELTFILTER && t->tcm_handle == 0) {
qdisc_lock_tree(dev);
*back = tp->next;
qdisc_unlock_tree(dev);
tfilter_notify(skb, n, tp, fh, RTM_DELTFILTER);
tcf_destroy(tp);
err = 0;
goto errout;
}
err = -ENOENT;
if (n->nlmsg_type != RTM_NEWTFILTER || !(n->nlmsg_flags&NLM_F_CREATE))
goto errout;
} else {
switch (n->nlmsg_type) {
case RTM_NEWTFILTER:
err = -EEXIST;
if (n->nlmsg_flags&NLM_F_EXCL)
goto errout;
break;
case RTM_DELTFILTER:
err = tp->ops->delete(tp, fh);
if (err == 0)
tfilter_notify(skb, n, tp, fh, RTM_DELTFILTER);
goto errout;
case RTM_GETTFILTER:
err = tfilter_notify(skb, n, tp, fh, RTM_NEWTFILTER);
goto errout;
default:
err = -EINVAL;
goto errout;
}
}
err = tp->ops->change(tp, cl, t->tcm_handle, tca, &fh);
if (err == 0)
tfilter_notify(skb, n, tp, fh, RTM_NEWTFILTER);
errout:
if (cl)
cops->put(q, cl);
if (err == -EAGAIN)
/* Replay the request. */
goto replay;
return err;
}
static int
tcf_fill_node(struct sk_buff *skb, struct tcf_proto *tp, unsigned long fh,
u32 pid, u32 seq, unsigned flags, int event)
{
struct tcmsg *tcm;
struct nlmsghdr *nlh;
unsigned char *b = skb->tail;
nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*tcm));
nlh->nlmsg_flags = flags;
tcm = NLMSG_DATA(nlh);
tcm->tcm_family = AF_UNSPEC;
tcm->tcm_ifindex = tp->q->dev->ifindex;
tcm->tcm_parent = tp->classid;
tcm->tcm_info = TC_H_MAKE(tp->prio, tp->protocol);
RTA_PUT(skb, TCA_KIND, IFNAMSIZ, tp->ops->kind);
tcm->tcm_handle = fh;
if (RTM_DELTFILTER != event) {
tcm->tcm_handle = 0;
if (tp->ops->dump && tp->ops->dump(tp, fh, skb, tcm) < 0)
goto rtattr_failure;
}
nlh->nlmsg_len = skb->tail - b;
return skb->len;
nlmsg_failure:
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int tfilter_notify(struct sk_buff *oskb, struct nlmsghdr *n,
struct tcf_proto *tp, unsigned long fh, int event)
{
struct sk_buff *skb;
u32 pid = oskb ? NETLINK_CB(oskb).pid : 0;
skb = alloc_skb(NLMSG_GOODSIZE, GFP_KERNEL);
if (!skb)
return -ENOBUFS;
if (tcf_fill_node(skb, tp, fh, pid, n->nlmsg_seq, 0, event) <= 0) {
kfree_skb(skb);
return -EINVAL;
}
return rtnetlink_send(skb, pid, RTMGRP_TC, n->nlmsg_flags&NLM_F_ECHO);
}
struct tcf_dump_args
{
struct tcf_walker w;
struct sk_buff *skb;
struct netlink_callback *cb;
};
static int tcf_node_dump(struct tcf_proto *tp, unsigned long n, struct tcf_walker *arg)
{
struct tcf_dump_args *a = (void*)arg;
return tcf_fill_node(a->skb, tp, n, NETLINK_CB(a->cb->skb).pid,
a->cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTFILTER);
}
static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb)
{
int t;
int s_t;
struct net_device *dev;
struct Qdisc *q;
struct tcf_proto *tp, **chain;
struct tcmsg *tcm = (struct tcmsg*)NLMSG_DATA(cb->nlh);
unsigned long cl = 0;
struct Qdisc_class_ops *cops;
struct tcf_dump_args arg;
if (cb->nlh->nlmsg_len < NLMSG_LENGTH(sizeof(*tcm)))
return skb->len;
if ((dev = dev_get_by_index(tcm->tcm_ifindex)) == NULL)
return skb->len;
read_lock_bh(&qdisc_tree_lock);
if (!tcm->tcm_parent)
q = dev->qdisc_sleeping;
else
q = qdisc_lookup(dev, TC_H_MAJ(tcm->tcm_parent));
if (!q)
goto out;
if ((cops = q->ops->cl_ops) == NULL)
goto errout;
if (TC_H_MIN(tcm->tcm_parent)) {
cl = cops->get(q, tcm->tcm_parent);
if (cl == 0)
goto errout;
}
chain = cops->tcf_chain(q, cl);
if (chain == NULL)
goto errout;
s_t = cb->args[0];
for (tp=*chain, t=0; tp; tp = tp->next, t++) {
if (t < s_t) continue;
if (TC_H_MAJ(tcm->tcm_info) &&
TC_H_MAJ(tcm->tcm_info) != tp->prio)
continue;
if (TC_H_MIN(tcm->tcm_info) &&
TC_H_MIN(tcm->tcm_info) != tp->protocol)
continue;
if (t > s_t)
memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(cb->args[0]));
if (cb->args[1] == 0) {
if (tcf_fill_node(skb, tp, 0, NETLINK_CB(cb->skb).pid,
cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWTFILTER) <= 0) {
break;
}
cb->args[1] = 1;
}
if (tp->ops->walk == NULL)
continue;
arg.w.fn = tcf_node_dump;
arg.skb = skb;
arg.cb = cb;
arg.w.stop = 0;
arg.w.skip = cb->args[1]-1;
arg.w.count = 0;
tp->ops->walk(tp, &arg.w);
cb->args[1] = arg.w.count+1;
if (arg.w.stop)
break;
}
cb->args[0] = t;
errout:
if (cl)
cops->put(q, cl);
out:
read_unlock_bh(&qdisc_tree_lock);
dev_put(dev);
return skb->len;
}
void
tcf_exts_destroy(struct tcf_proto *tp, struct tcf_exts *exts)
{
#ifdef CONFIG_NET_CLS_ACT
if (exts->action) {
tcf_action_destroy(exts->action, TCA_ACT_UNBIND);
exts->action = NULL;
}
#elif defined CONFIG_NET_CLS_POLICE
if (exts->police) {
tcf_police_release(exts->police, TCA_ACT_UNBIND);
exts->police = NULL;
}
#endif
}
int
tcf_exts_validate(struct tcf_proto *tp, struct rtattr **tb,
struct rtattr *rate_tlv, struct tcf_exts *exts,
struct tcf_ext_map *map)
{
memset(exts, 0, sizeof(*exts));
#ifdef CONFIG_NET_CLS_ACT
{
int err;
struct tc_action *act;
if (map->police && tb[map->police-1]) {
act = tcf_action_init_1(tb[map->police-1], rate_tlv, "police",
TCA_ACT_NOREPLACE, TCA_ACT_BIND, &err);
if (act == NULL)
return err;
act->type = TCA_OLD_COMPAT;
exts->action = act;
} else if (map->action && tb[map->action-1]) {
act = tcf_action_init(tb[map->action-1], rate_tlv, NULL,
TCA_ACT_NOREPLACE, TCA_ACT_BIND, &err);
if (act == NULL)
return err;
exts->action = act;
}
}
#elif defined CONFIG_NET_CLS_POLICE
if (map->police && tb[map->police-1]) {
struct tcf_police *p;
p = tcf_police_locate(tb[map->police-1], rate_tlv);
if (p == NULL)
return -EINVAL;
exts->police = p;
} else if (map->action && tb[map->action-1])
return -EOPNOTSUPP;
#else
if ((map->action && tb[map->action-1]) ||
(map->police && tb[map->police-1]))
return -EOPNOTSUPP;
#endif
return 0;
}
void
tcf_exts_change(struct tcf_proto *tp, struct tcf_exts *dst,
struct tcf_exts *src)
{
#ifdef CONFIG_NET_CLS_ACT
if (src->action) {
struct tc_action *act;
tcf_tree_lock(tp);
act = xchg(&dst->action, src->action);
tcf_tree_unlock(tp);
if (act)
tcf_action_destroy(act, TCA_ACT_UNBIND);
}
#elif defined CONFIG_NET_CLS_POLICE
if (src->police) {
struct tcf_police *p;
tcf_tree_lock(tp);
p = xchg(&dst->police, src->police);
tcf_tree_unlock(tp);
if (p)
tcf_police_release(p, TCA_ACT_UNBIND);
}
#endif
}
int
tcf_exts_dump(struct sk_buff *skb, struct tcf_exts *exts,
struct tcf_ext_map *map)
{
#ifdef CONFIG_NET_CLS_ACT
if (map->action && exts->action) {
/*
* again for backward compatible mode - we want
* to work with both old and new modes of entering
* tc data even if iproute2 was newer - jhs
*/
struct rtattr * p_rta = (struct rtattr*) skb->tail;
if (exts->action->type != TCA_OLD_COMPAT) {
RTA_PUT(skb, map->action, 0, NULL);
if (tcf_action_dump(skb, exts->action, 0, 0) < 0)
goto rtattr_failure;
p_rta->rta_len = skb->tail - (u8*)p_rta;
} else if (map->police) {
RTA_PUT(skb, map->police, 0, NULL);
if (tcf_action_dump_old(skb, exts->action, 0, 0) < 0)
goto rtattr_failure;
p_rta->rta_len = skb->tail - (u8*)p_rta;
}
}
#elif defined CONFIG_NET_CLS_POLICE
if (map->police && exts->police) {
struct rtattr * p_rta = (struct rtattr*) skb->tail;
RTA_PUT(skb, map->police, 0, NULL);
if (tcf_police_dump(skb, exts->police) < 0)
goto rtattr_failure;
p_rta->rta_len = skb->tail - (u8*)p_rta;
}
#endif
return 0;
rtattr_failure: __attribute__ ((unused))
return -1;
}
int
tcf_exts_dump_stats(struct sk_buff *skb, struct tcf_exts *exts,
struct tcf_ext_map *map)
{
#ifdef CONFIG_NET_CLS_ACT
if (exts->action)
if (tcf_action_copy_stats(skb, exts->action, 1) < 0)
goto rtattr_failure;
#elif defined CONFIG_NET_CLS_POLICE
if (exts->police)
if (tcf_police_dump_stats(skb, exts->police) < 0)
goto rtattr_failure;
#endif
return 0;
rtattr_failure: __attribute__ ((unused))
return -1;
}
static int __init tc_filter_init(void)
{
struct rtnetlink_link *link_p = rtnetlink_links[PF_UNSPEC];
/* Setup rtnetlink links. It is made here to avoid
exporting large number of public symbols.
*/
if (link_p) {
link_p[RTM_NEWTFILTER-RTM_BASE].doit = tc_ctl_tfilter;
link_p[RTM_DELTFILTER-RTM_BASE].doit = tc_ctl_tfilter;
link_p[RTM_GETTFILTER-RTM_BASE].doit = tc_ctl_tfilter;
link_p[RTM_GETTFILTER-RTM_BASE].dumpit = tc_dump_tfilter;
}
return 0;
}
subsys_initcall(tc_filter_init);
EXPORT_SYMBOL(register_tcf_proto_ops);
EXPORT_SYMBOL(unregister_tcf_proto_ops);
EXPORT_SYMBOL(tcf_exts_validate);
EXPORT_SYMBOL(tcf_exts_destroy);
EXPORT_SYMBOL(tcf_exts_change);
EXPORT_SYMBOL(tcf_exts_dump);
EXPORT_SYMBOL(tcf_exts_dump_stats);

303
net/sched/cls_basic.c Normal file
View File

@@ -0,0 +1,303 @@
/*
* net/sched/cls_basic.c Basic Packet Classifier.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Thomas Graf <tgraf@suug.ch>
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/errno.h>
#include <linux/rtnetlink.h>
#include <linux/skbuff.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>
struct basic_head
{
u32 hgenerator;
struct list_head flist;
};
struct basic_filter
{
u32 handle;
struct tcf_exts exts;
struct tcf_ematch_tree ematches;
struct tcf_result res;
struct list_head link;
};
static struct tcf_ext_map basic_ext_map = {
.action = TCA_BASIC_ACT,
.police = TCA_BASIC_POLICE
};
static int basic_classify(struct sk_buff *skb, struct tcf_proto *tp,
struct tcf_result *res)
{
int r;
struct basic_head *head = (struct basic_head *) tp->root;
struct basic_filter *f;
list_for_each_entry(f, &head->flist, link) {
if (!tcf_em_tree_match(skb, &f->ematches, NULL))
continue;
*res = f->res;
r = tcf_exts_exec(skb, &f->exts, res);
if (r < 0)
continue;
return r;
}
return -1;
}
static unsigned long basic_get(struct tcf_proto *tp, u32 handle)
{
unsigned long l = 0UL;
struct basic_head *head = (struct basic_head *) tp->root;
struct basic_filter *f;
if (head == NULL)
return 0UL;
list_for_each_entry(f, &head->flist, link)
if (f->handle == handle)
l = (unsigned long) f;
return l;
}
static void basic_put(struct tcf_proto *tp, unsigned long f)
{
}
static int basic_init(struct tcf_proto *tp)
{
return 0;
}
static inline void basic_delete_filter(struct tcf_proto *tp,
struct basic_filter *f)
{
tcf_unbind_filter(tp, &f->res);
tcf_exts_destroy(tp, &f->exts);
tcf_em_tree_destroy(tp, &f->ematches);
kfree(f);
}
static void basic_destroy(struct tcf_proto *tp)
{
struct basic_head *head = (struct basic_head *) xchg(&tp->root, NULL);
struct basic_filter *f, *n;
list_for_each_entry_safe(f, n, &head->flist, link) {
list_del(&f->link);
basic_delete_filter(tp, f);
}
}
static int basic_delete(struct tcf_proto *tp, unsigned long arg)
{
struct basic_head *head = (struct basic_head *) tp->root;
struct basic_filter *t, *f = (struct basic_filter *) arg;
list_for_each_entry(t, &head->flist, link)
if (t == f) {
tcf_tree_lock(tp);
list_del(&t->link);
tcf_tree_unlock(tp);
basic_delete_filter(tp, t);
return 0;
}
return -ENOENT;
}
static inline int basic_set_parms(struct tcf_proto *tp, struct basic_filter *f,
unsigned long base, struct rtattr **tb,
struct rtattr *est)
{
int err = -EINVAL;
struct tcf_exts e;
struct tcf_ematch_tree t;
if (tb[TCA_BASIC_CLASSID-1])
if (RTA_PAYLOAD(tb[TCA_BASIC_CLASSID-1]) < sizeof(u32))
return err;
err = tcf_exts_validate(tp, tb, est, &e, &basic_ext_map);
if (err < 0)
return err;
err = tcf_em_tree_validate(tp, tb[TCA_BASIC_EMATCHES-1], &t);
if (err < 0)
goto errout;
if (tb[TCA_BASIC_CLASSID-1]) {
f->res.classid = *(u32*)RTA_DATA(tb[TCA_BASIC_CLASSID-1]);
tcf_bind_filter(tp, &f->res, base);
}
tcf_exts_change(tp, &f->exts, &e);
tcf_em_tree_change(tp, &f->ematches, &t);
return 0;
errout:
tcf_exts_destroy(tp, &e);
return err;
}
static int basic_change(struct tcf_proto *tp, unsigned long base, u32 handle,
struct rtattr **tca, unsigned long *arg)
{
int err = -EINVAL;
struct basic_head *head = (struct basic_head *) tp->root;
struct rtattr *tb[TCA_BASIC_MAX];
struct basic_filter *f = (struct basic_filter *) *arg;
if (tca[TCA_OPTIONS-1] == NULL)
return -EINVAL;
if (rtattr_parse_nested(tb, TCA_BASIC_MAX, tca[TCA_OPTIONS-1]) < 0)
return -EINVAL;
if (f != NULL) {
if (handle && f->handle != handle)
return -EINVAL;
return basic_set_parms(tp, f, base, tb, tca[TCA_RATE-1]);
}
err = -ENOBUFS;
if (head == NULL) {
head = kmalloc(sizeof(*head), GFP_KERNEL);
if (head == NULL)
goto errout;
memset(head, 0, sizeof(*head));
INIT_LIST_HEAD(&head->flist);
tp->root = head;
}
f = kmalloc(sizeof(*f), GFP_KERNEL);
if (f == NULL)
goto errout;
memset(f, 0, sizeof(*f));
err = -EINVAL;
if (handle)
f->handle = handle;
else {
int i = 0x80000000;
do {
if (++head->hgenerator == 0x7FFFFFFF)
head->hgenerator = 1;
} while (--i > 0 && basic_get(tp, head->hgenerator));
if (i <= 0) {
printk(KERN_ERR "Insufficient number of handles\n");
goto errout;
}
f->handle = head->hgenerator;
}
err = basic_set_parms(tp, f, base, tb, tca[TCA_RATE-1]);
if (err < 0)
goto errout;
tcf_tree_lock(tp);
list_add(&f->link, &head->flist);
tcf_tree_unlock(tp);
*arg = (unsigned long) f;
return 0;
errout:
if (*arg == 0UL && f)
kfree(f);
return err;
}
static void basic_walk(struct tcf_proto *tp, struct tcf_walker *arg)
{
struct basic_head *head = (struct basic_head *) tp->root;
struct basic_filter *f;
list_for_each_entry(f, &head->flist, link) {
if (arg->count < arg->skip)
goto skip;
if (arg->fn(tp, (unsigned long) f, arg) < 0) {
arg->stop = 1;
break;
}
skip:
arg->count++;
}
}
static int basic_dump(struct tcf_proto *tp, unsigned long fh,
struct sk_buff *skb, struct tcmsg *t)
{
struct basic_filter *f = (struct basic_filter *) fh;
unsigned char *b = skb->tail;
struct rtattr *rta;
if (f == NULL)
return skb->len;
t->tcm_handle = f->handle;
rta = (struct rtattr *) b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
if (tcf_exts_dump(skb, &f->exts, &basic_ext_map) < 0 ||
tcf_em_tree_dump(skb, &f->ematches, TCA_BASIC_EMATCHES) < 0)
goto rtattr_failure;
rta->rta_len = (skb->tail - b);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tcf_proto_ops cls_basic_ops = {
.kind = "basic",
.classify = basic_classify,
.init = basic_init,
.destroy = basic_destroy,
.get = basic_get,
.put = basic_put,
.change = basic_change,
.delete = basic_delete,
.walk = basic_walk,
.dump = basic_dump,
.owner = THIS_MODULE,
};
static int __init init_basic(void)
{
return register_tcf_proto_ops(&cls_basic_ops);
}
static void __exit exit_basic(void)
{
unregister_tcf_proto_ops(&cls_basic_ops);
}
module_init(init_basic)
module_exit(exit_basic)
MODULE_LICENSE("GPL");

378
net/sched/cls_fw.c Normal file
View File

@@ -0,0 +1,378 @@
/*
* net/sched/cls_fw.c Classifier mapping ipchains' fwmark to traffic class.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
* Changes:
* Karlis Peisenieks <karlis@mt.lv> : 990415 : fw_walk off by one
* Karlis Peisenieks <karlis@mt.lv> : 990415 : fw_delete killed all the filter (and kernel).
* Alex <alex@pilotsoft.com> : 2004xxyy: Added Action extension
*
* JHS: We should remove the CONFIG_NET_CLS_IND from here
* eventually when the meta match extension is made available
*
*/
#include <linux/config.h>
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <linux/netfilter.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>
struct fw_head
{
struct fw_filter *ht[256];
};
struct fw_filter
{
struct fw_filter *next;
u32 id;
struct tcf_result res;
#ifdef CONFIG_NET_CLS_IND
char indev[IFNAMSIZ];
#endif /* CONFIG_NET_CLS_IND */
struct tcf_exts exts;
};
static struct tcf_ext_map fw_ext_map = {
.action = TCA_FW_ACT,
.police = TCA_FW_POLICE
};
static __inline__ int fw_hash(u32 handle)
{
return handle&0xFF;
}
static int fw_classify(struct sk_buff *skb, struct tcf_proto *tp,
struct tcf_result *res)
{
struct fw_head *head = (struct fw_head*)tp->root;
struct fw_filter *f;
int r;
#ifdef CONFIG_NETFILTER
u32 id = skb->nfmark;
#else
u32 id = 0;
#endif
if (head != NULL) {
for (f=head->ht[fw_hash(id)]; f; f=f->next) {
if (f->id == id) {
*res = f->res;
#ifdef CONFIG_NET_CLS_IND
if (!tcf_match_indev(skb, f->indev))
continue;
#endif /* CONFIG_NET_CLS_IND */
r = tcf_exts_exec(skb, &f->exts, res);
if (r < 0)
continue;
return r;
}
}
} else {
/* old method */
if (id && (TC_H_MAJ(id) == 0 || !(TC_H_MAJ(id^tp->q->handle)))) {
res->classid = id;
res->class = 0;
return 0;
}
}
return -1;
}
static unsigned long fw_get(struct tcf_proto *tp, u32 handle)
{
struct fw_head *head = (struct fw_head*)tp->root;
struct fw_filter *f;
if (head == NULL)
return 0;
for (f=head->ht[fw_hash(handle)]; f; f=f->next) {
if (f->id == handle)
return (unsigned long)f;
}
return 0;
}
static void fw_put(struct tcf_proto *tp, unsigned long f)
{
}
static int fw_init(struct tcf_proto *tp)
{
return 0;
}
static inline void
fw_delete_filter(struct tcf_proto *tp, struct fw_filter *f)
{
tcf_unbind_filter(tp, &f->res);
tcf_exts_destroy(tp, &f->exts);
kfree(f);
}
static void fw_destroy(struct tcf_proto *tp)
{
struct fw_head *head = (struct fw_head*)xchg(&tp->root, NULL);
struct fw_filter *f;
int h;
if (head == NULL)
return;
for (h=0; h<256; h++) {
while ((f=head->ht[h]) != NULL) {
head->ht[h] = f->next;
fw_delete_filter(tp, f);
}
}
kfree(head);
}
static int fw_delete(struct tcf_proto *tp, unsigned long arg)
{
struct fw_head *head = (struct fw_head*)tp->root;
struct fw_filter *f = (struct fw_filter*)arg;
struct fw_filter **fp;
if (head == NULL || f == NULL)
goto out;
for (fp=&head->ht[fw_hash(f->id)]; *fp; fp = &(*fp)->next) {
if (*fp == f) {
tcf_tree_lock(tp);
*fp = f->next;
tcf_tree_unlock(tp);
fw_delete_filter(tp, f);
return 0;
}
}
out:
return -EINVAL;
}
static int
fw_change_attrs(struct tcf_proto *tp, struct fw_filter *f,
struct rtattr **tb, struct rtattr **tca, unsigned long base)
{
struct tcf_exts e;
int err;
err = tcf_exts_validate(tp, tb, tca[TCA_RATE-1], &e, &fw_ext_map);
if (err < 0)
return err;
err = -EINVAL;
if (tb[TCA_FW_CLASSID-1]) {
if (RTA_PAYLOAD(tb[TCA_FW_CLASSID-1]) != sizeof(u32))
goto errout;
f->res.classid = *(u32*)RTA_DATA(tb[TCA_FW_CLASSID-1]);
tcf_bind_filter(tp, &f->res, base);
}
#ifdef CONFIG_NET_CLS_IND
if (tb[TCA_FW_INDEV-1]) {
err = tcf_change_indev(tp, f->indev, tb[TCA_FW_INDEV-1]);
if (err < 0)
goto errout;
}
#endif /* CONFIG_NET_CLS_IND */
tcf_exts_change(tp, &f->exts, &e);
return 0;
errout:
tcf_exts_destroy(tp, &e);
return err;
}
static int fw_change(struct tcf_proto *tp, unsigned long base,
u32 handle,
struct rtattr **tca,
unsigned long *arg)
{
struct fw_head *head = (struct fw_head*)tp->root;
struct fw_filter *f = (struct fw_filter *) *arg;
struct rtattr *opt = tca[TCA_OPTIONS-1];
struct rtattr *tb[TCA_FW_MAX];
int err;
if (!opt)
return handle ? -EINVAL : 0;
if (rtattr_parse_nested(tb, TCA_FW_MAX, opt) < 0)
return -EINVAL;
if (f != NULL) {
if (f->id != handle && handle)
return -EINVAL;
return fw_change_attrs(tp, f, tb, tca, base);
}
if (!handle)
return -EINVAL;
if (head == NULL) {
head = kmalloc(sizeof(struct fw_head), GFP_KERNEL);
if (head == NULL)
return -ENOBUFS;
memset(head, 0, sizeof(*head));
tcf_tree_lock(tp);
tp->root = head;
tcf_tree_unlock(tp);
}
f = kmalloc(sizeof(struct fw_filter), GFP_KERNEL);
if (f == NULL)
return -ENOBUFS;
memset(f, 0, sizeof(*f));
f->id = handle;
err = fw_change_attrs(tp, f, tb, tca, base);
if (err < 0)
goto errout;
f->next = head->ht[fw_hash(handle)];
tcf_tree_lock(tp);
head->ht[fw_hash(handle)] = f;
tcf_tree_unlock(tp);
*arg = (unsigned long)f;
return 0;
errout:
if (f)
kfree(f);
return err;
}
static void fw_walk(struct tcf_proto *tp, struct tcf_walker *arg)
{
struct fw_head *head = (struct fw_head*)tp->root;
int h;
if (head == NULL)
arg->stop = 1;
if (arg->stop)
return;
for (h = 0; h < 256; h++) {
struct fw_filter *f;
for (f = head->ht[h]; f; f = f->next) {
if (arg->count < arg->skip) {
arg->count++;
continue;
}
if (arg->fn(tp, (unsigned long)f, arg) < 0) {
arg->stop = 1;
return;
}
arg->count++;
}
}
}
static int fw_dump(struct tcf_proto *tp, unsigned long fh,
struct sk_buff *skb, struct tcmsg *t)
{
struct fw_filter *f = (struct fw_filter*)fh;
unsigned char *b = skb->tail;
struct rtattr *rta;
if (f == NULL)
return skb->len;
t->tcm_handle = f->id;
if (!f->res.classid && !tcf_exts_is_available(&f->exts))
return skb->len;
rta = (struct rtattr*)b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
if (f->res.classid)
RTA_PUT(skb, TCA_FW_CLASSID, 4, &f->res.classid);
#ifdef CONFIG_NET_CLS_IND
if (strlen(f->indev))
RTA_PUT(skb, TCA_FW_INDEV, IFNAMSIZ, f->indev);
#endif /* CONFIG_NET_CLS_IND */
if (tcf_exts_dump(skb, &f->exts, &fw_ext_map) < 0)
goto rtattr_failure;
rta->rta_len = skb->tail - b;
if (tcf_exts_dump_stats(skb, &f->exts, &fw_ext_map) < 0)
goto rtattr_failure;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tcf_proto_ops cls_fw_ops = {
.next = NULL,
.kind = "fw",
.classify = fw_classify,
.init = fw_init,
.destroy = fw_destroy,
.get = fw_get,
.put = fw_put,
.change = fw_change,
.delete = fw_delete,
.walk = fw_walk,
.dump = fw_dump,
.owner = THIS_MODULE,
};
static int __init init_fw(void)
{
return register_tcf_proto_ops(&cls_fw_ops);
}
static void __exit exit_fw(void)
{
unregister_tcf_proto_ops(&cls_fw_ops);
}
module_init(init_fw)
module_exit(exit_fw)
MODULE_LICENSE("GPL");

639
net/sched/cls_route.c Normal file
View File

@@ -0,0 +1,639 @@
/*
* net/sched/cls_route.c ROUTE4 classifier.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
#include <linux/module.h>
#include <linux/config.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>
/*
1. For now we assume that route tags < 256.
It allows to use direct table lookups, instead of hash tables.
2. For now we assume that "from TAG" and "fromdev DEV" statements
are mutually exclusive.
3. "to TAG from ANY" has higher priority, than "to ANY from XXX"
*/
struct route4_fastmap
{
struct route4_filter *filter;
u32 id;
int iif;
};
struct route4_head
{
struct route4_fastmap fastmap[16];
struct route4_bucket *table[256+1];
};
struct route4_bucket
{
/* 16 FROM buckets + 16 IIF buckets + 1 wildcard bucket */
struct route4_filter *ht[16+16+1];
};
struct route4_filter
{
struct route4_filter *next;
u32 id;
int iif;
struct tcf_result res;
struct tcf_exts exts;
u32 handle;
struct route4_bucket *bkt;
};
#define ROUTE4_FAILURE ((struct route4_filter*)(-1L))
static struct tcf_ext_map route_ext_map = {
.police = TCA_ROUTE4_POLICE,
.action = TCA_ROUTE4_ACT
};
static __inline__ int route4_fastmap_hash(u32 id, int iif)
{
return id&0xF;
}
static inline
void route4_reset_fastmap(struct net_device *dev, struct route4_head *head, u32 id)
{
spin_lock_bh(&dev->queue_lock);
memset(head->fastmap, 0, sizeof(head->fastmap));
spin_unlock_bh(&dev->queue_lock);
}
static void __inline__
route4_set_fastmap(struct route4_head *head, u32 id, int iif,
struct route4_filter *f)
{
int h = route4_fastmap_hash(id, iif);
head->fastmap[h].id = id;
head->fastmap[h].iif = iif;
head->fastmap[h].filter = f;
}
static __inline__ int route4_hash_to(u32 id)
{
return id&0xFF;
}
static __inline__ int route4_hash_from(u32 id)
{
return (id>>16)&0xF;
}
static __inline__ int route4_hash_iif(int iif)
{
return 16 + ((iif>>16)&0xF);
}
static __inline__ int route4_hash_wild(void)
{
return 32;
}
#define ROUTE4_APPLY_RESULT() \
{ \
*res = f->res; \
if (tcf_exts_is_available(&f->exts)) { \
int r = tcf_exts_exec(skb, &f->exts, res); \
if (r < 0) { \
dont_cache = 1; \
continue; \
} \
return r; \
} else if (!dont_cache) \
route4_set_fastmap(head, id, iif, f); \
return 0; \
}
static int route4_classify(struct sk_buff *skb, struct tcf_proto *tp,
struct tcf_result *res)
{
struct route4_head *head = (struct route4_head*)tp->root;
struct dst_entry *dst;
struct route4_bucket *b;
struct route4_filter *f;
u32 id, h;
int iif, dont_cache = 0;
if ((dst = skb->dst) == NULL)
goto failure;
id = dst->tclassid;
if (head == NULL)
goto old_method;
iif = ((struct rtable*)dst)->fl.iif;
h = route4_fastmap_hash(id, iif);
if (id == head->fastmap[h].id &&
iif == head->fastmap[h].iif &&
(f = head->fastmap[h].filter) != NULL) {
if (f == ROUTE4_FAILURE)
goto failure;
*res = f->res;
return 0;
}
h = route4_hash_to(id);
restart:
if ((b = head->table[h]) != NULL) {
for (f = b->ht[route4_hash_from(id)]; f; f = f->next)
if (f->id == id)
ROUTE4_APPLY_RESULT();
for (f = b->ht[route4_hash_iif(iif)]; f; f = f->next)
if (f->iif == iif)
ROUTE4_APPLY_RESULT();
for (f = b->ht[route4_hash_wild()]; f; f = f->next)
ROUTE4_APPLY_RESULT();
}
if (h < 256) {
h = 256;
id &= ~0xFFFF;
goto restart;
}
if (!dont_cache)
route4_set_fastmap(head, id, iif, ROUTE4_FAILURE);
failure:
return -1;
old_method:
if (id && (TC_H_MAJ(id) == 0 ||
!(TC_H_MAJ(id^tp->q->handle)))) {
res->classid = id;
res->class = 0;
return 0;
}
return -1;
}
static inline u32 to_hash(u32 id)
{
u32 h = id&0xFF;
if (id&0x8000)
h += 256;
return h;
}
static inline u32 from_hash(u32 id)
{
id &= 0xFFFF;
if (id == 0xFFFF)
return 32;
if (!(id & 0x8000)) {
if (id > 255)
return 256;
return id&0xF;
}
return 16 + (id&0xF);
}
static unsigned long route4_get(struct tcf_proto *tp, u32 handle)
{
struct route4_head *head = (struct route4_head*)tp->root;
struct route4_bucket *b;
struct route4_filter *f;
unsigned h1, h2;
if (!head)
return 0;
h1 = to_hash(handle);
if (h1 > 256)
return 0;
h2 = from_hash(handle>>16);
if (h2 > 32)
return 0;
if ((b = head->table[h1]) != NULL) {
for (f = b->ht[h2]; f; f = f->next)
if (f->handle == handle)
return (unsigned long)f;
}
return 0;
}
static void route4_put(struct tcf_proto *tp, unsigned long f)
{
}
static int route4_init(struct tcf_proto *tp)
{
return 0;
}
static inline void
route4_delete_filter(struct tcf_proto *tp, struct route4_filter *f)
{
tcf_unbind_filter(tp, &f->res);
tcf_exts_destroy(tp, &f->exts);
kfree(f);
}
static void route4_destroy(struct tcf_proto *tp)
{
struct route4_head *head = xchg(&tp->root, NULL);
int h1, h2;
if (head == NULL)
return;
for (h1=0; h1<=256; h1++) {
struct route4_bucket *b;
if ((b = head->table[h1]) != NULL) {
for (h2=0; h2<=32; h2++) {
struct route4_filter *f;
while ((f = b->ht[h2]) != NULL) {
b->ht[h2] = f->next;
route4_delete_filter(tp, f);
}
}
kfree(b);
}
}
kfree(head);
}
static int route4_delete(struct tcf_proto *tp, unsigned long arg)
{
struct route4_head *head = (struct route4_head*)tp->root;
struct route4_filter **fp, *f = (struct route4_filter*)arg;
unsigned h = 0;
struct route4_bucket *b;
int i;
if (!head || !f)
return -EINVAL;
h = f->handle;
b = f->bkt;
for (fp = &b->ht[from_hash(h>>16)]; *fp; fp = &(*fp)->next) {
if (*fp == f) {
tcf_tree_lock(tp);
*fp = f->next;
tcf_tree_unlock(tp);
route4_reset_fastmap(tp->q->dev, head, f->id);
route4_delete_filter(tp, f);
/* Strip tree */
for (i=0; i<=32; i++)
if (b->ht[i])
return 0;
/* OK, session has no flows */
tcf_tree_lock(tp);
head->table[to_hash(h)] = NULL;
tcf_tree_unlock(tp);
kfree(b);
return 0;
}
}
return 0;
}
static int route4_set_parms(struct tcf_proto *tp, unsigned long base,
struct route4_filter *f, u32 handle, struct route4_head *head,
struct rtattr **tb, struct rtattr *est, int new)
{
int err;
u32 id = 0, to = 0, nhandle = 0x8000;
struct route4_filter *fp;
unsigned int h1;
struct route4_bucket *b;
struct tcf_exts e;
err = tcf_exts_validate(tp, tb, est, &e, &route_ext_map);
if (err < 0)
return err;
err = -EINVAL;
if (tb[TCA_ROUTE4_CLASSID-1])
if (RTA_PAYLOAD(tb[TCA_ROUTE4_CLASSID-1]) < sizeof(u32))
goto errout;
if (tb[TCA_ROUTE4_TO-1]) {
if (new && handle & 0x8000)
goto errout;
if (RTA_PAYLOAD(tb[TCA_ROUTE4_TO-1]) < sizeof(u32))
goto errout;
to = *(u32*)RTA_DATA(tb[TCA_ROUTE4_TO-1]);
if (to > 0xFF)
goto errout;
nhandle = to;
}
if (tb[TCA_ROUTE4_FROM-1]) {
if (tb[TCA_ROUTE4_IIF-1])
goto errout;
if (RTA_PAYLOAD(tb[TCA_ROUTE4_FROM-1]) < sizeof(u32))
goto errout;
id = *(u32*)RTA_DATA(tb[TCA_ROUTE4_FROM-1]);
if (id > 0xFF)
goto errout;
nhandle |= id << 16;
} else if (tb[TCA_ROUTE4_IIF-1]) {
if (RTA_PAYLOAD(tb[TCA_ROUTE4_IIF-1]) < sizeof(u32))
goto errout;
id = *(u32*)RTA_DATA(tb[TCA_ROUTE4_IIF-1]);
if (id > 0x7FFF)
goto errout;
nhandle |= (id | 0x8000) << 16;
} else
nhandle |= 0xFFFF << 16;
if (handle && new) {
nhandle |= handle & 0x7F00;
if (nhandle != handle)
goto errout;
}
h1 = to_hash(nhandle);
if ((b = head->table[h1]) == NULL) {
err = -ENOBUFS;
b = kmalloc(sizeof(struct route4_bucket), GFP_KERNEL);
if (b == NULL)
goto errout;
memset(b, 0, sizeof(*b));
tcf_tree_lock(tp);
head->table[h1] = b;
tcf_tree_unlock(tp);
} else {
unsigned int h2 = from_hash(nhandle >> 16);
err = -EEXIST;
for (fp = b->ht[h2]; fp; fp = fp->next)
if (fp->handle == f->handle)
goto errout;
}
tcf_tree_lock(tp);
if (tb[TCA_ROUTE4_TO-1])
f->id = to;
if (tb[TCA_ROUTE4_FROM-1])
f->id = to | id<<16;
else if (tb[TCA_ROUTE4_IIF-1])
f->iif = id;
f->handle = nhandle;
f->bkt = b;
tcf_tree_unlock(tp);
if (tb[TCA_ROUTE4_CLASSID-1]) {
f->res.classid = *(u32*)RTA_DATA(tb[TCA_ROUTE4_CLASSID-1]);
tcf_bind_filter(tp, &f->res, base);
}
tcf_exts_change(tp, &f->exts, &e);
return 0;
errout:
tcf_exts_destroy(tp, &e);
return err;
}
static int route4_change(struct tcf_proto *tp, unsigned long base,
u32 handle,
struct rtattr **tca,
unsigned long *arg)
{
struct route4_head *head = tp->root;
struct route4_filter *f, *f1, **fp;
struct route4_bucket *b;
struct rtattr *opt = tca[TCA_OPTIONS-1];
struct rtattr *tb[TCA_ROUTE4_MAX];
unsigned int h, th;
u32 old_handle = 0;
int err;
if (opt == NULL)
return handle ? -EINVAL : 0;
if (rtattr_parse_nested(tb, TCA_ROUTE4_MAX, opt) < 0)
return -EINVAL;
if ((f = (struct route4_filter*)*arg) != NULL) {
if (f->handle != handle && handle)
return -EINVAL;
if (f->bkt)
old_handle = f->handle;
err = route4_set_parms(tp, base, f, handle, head, tb,
tca[TCA_RATE-1], 0);
if (err < 0)
return err;
goto reinsert;
}
err = -ENOBUFS;
if (head == NULL) {
head = kmalloc(sizeof(struct route4_head), GFP_KERNEL);
if (head == NULL)
goto errout;
memset(head, 0, sizeof(struct route4_head));
tcf_tree_lock(tp);
tp->root = head;
tcf_tree_unlock(tp);
}
f = kmalloc(sizeof(struct route4_filter), GFP_KERNEL);
if (f == NULL)
goto errout;
memset(f, 0, sizeof(*f));
err = route4_set_parms(tp, base, f, handle, head, tb,
tca[TCA_RATE-1], 1);
if (err < 0)
goto errout;
reinsert:
h = from_hash(f->handle >> 16);
for (fp = &f->bkt->ht[h]; (f1=*fp) != NULL; fp = &f1->next)
if (f->handle < f1->handle)
break;
f->next = f1;
tcf_tree_lock(tp);
*fp = f;
if (old_handle && f->handle != old_handle) {
th = to_hash(old_handle);
h = from_hash(old_handle >> 16);
if ((b = head->table[th]) != NULL) {
for (fp = &b->ht[h]; *fp; fp = &(*fp)->next) {
if (*fp == f) {
*fp = f->next;
break;
}
}
}
}
tcf_tree_unlock(tp);
route4_reset_fastmap(tp->q->dev, head, f->id);
*arg = (unsigned long)f;
return 0;
errout:
if (f)
kfree(f);
return err;
}
static void route4_walk(struct tcf_proto *tp, struct tcf_walker *arg)
{
struct route4_head *head = tp->root;
unsigned h, h1;
if (head == NULL)
arg->stop = 1;
if (arg->stop)
return;
for (h = 0; h <= 256; h++) {
struct route4_bucket *b = head->table[h];
if (b) {
for (h1 = 0; h1 <= 32; h1++) {
struct route4_filter *f;
for (f = b->ht[h1]; f; f = f->next) {
if (arg->count < arg->skip) {
arg->count++;
continue;
}
if (arg->fn(tp, (unsigned long)f, arg) < 0) {
arg->stop = 1;
return;
}
arg->count++;
}
}
}
}
}
static int route4_dump(struct tcf_proto *tp, unsigned long fh,
struct sk_buff *skb, struct tcmsg *t)
{
struct route4_filter *f = (struct route4_filter*)fh;
unsigned char *b = skb->tail;
struct rtattr *rta;
u32 id;
if (f == NULL)
return skb->len;
t->tcm_handle = f->handle;
rta = (struct rtattr*)b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
if (!(f->handle&0x8000)) {
id = f->id&0xFF;
RTA_PUT(skb, TCA_ROUTE4_TO, sizeof(id), &id);
}
if (f->handle&0x80000000) {
if ((f->handle>>16) != 0xFFFF)
RTA_PUT(skb, TCA_ROUTE4_IIF, sizeof(f->iif), &f->iif);
} else {
id = f->id>>16;
RTA_PUT(skb, TCA_ROUTE4_FROM, sizeof(id), &id);
}
if (f->res.classid)
RTA_PUT(skb, TCA_ROUTE4_CLASSID, 4, &f->res.classid);
if (tcf_exts_dump(skb, &f->exts, &route_ext_map) < 0)
goto rtattr_failure;
rta->rta_len = skb->tail - b;
if (tcf_exts_dump_stats(skb, &f->exts, &route_ext_map) < 0)
goto rtattr_failure;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tcf_proto_ops cls_route4_ops = {
.next = NULL,
.kind = "route",
.classify = route4_classify,
.init = route4_init,
.destroy = route4_destroy,
.get = route4_get,
.put = route4_put,
.change = route4_change,
.delete = route4_delete,
.walk = route4_walk,
.dump = route4_dump,
.owner = THIS_MODULE,
};
static int __init init_route4(void)
{
return register_tcf_proto_ops(&cls_route4_ops);
}
static void __exit exit_route4(void)
{
unregister_tcf_proto_ops(&cls_route4_ops);
}
module_init(init_route4)
module_exit(exit_route4)
MODULE_LICENSE("GPL");

43
net/sched/cls_rsvp.c Normal file
View File

@@ -0,0 +1,43 @@
/*
* net/sched/cls_rsvp.c Special RSVP packet classifier for IPv4.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>
#define RSVP_DST_LEN 1
#define RSVP_ID "rsvp"
#define RSVP_OPS cls_rsvp_ops
#include "cls_rsvp.h"
MODULE_LICENSE("GPL");

667
net/sched/cls_rsvp.h Normal file
View File

@@ -0,0 +1,667 @@
/*
* net/sched/cls_rsvp.h Template file for RSVPv[46] classifiers.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
/*
Comparing to general packet classification problem,
RSVP needs only sevaral relatively simple rules:
* (dst, protocol) are always specified,
so that we are able to hash them.
* src may be exact, or may be wildcard, so that
we can keep a hash table plus one wildcard entry.
* source port (or flow label) is important only if src is given.
IMPLEMENTATION.
We use a two level hash table: The top level is keyed by
destination address and protocol ID, every bucket contains a list
of "rsvp sessions", identified by destination address, protocol and
DPI(="Destination Port ID"): triple (key, mask, offset).
Every bucket has a smaller hash table keyed by source address
(cf. RSVP flowspec) and one wildcard entry for wildcard reservations.
Every bucket is again a list of "RSVP flows", selected by
source address and SPI(="Source Port ID" here rather than
"security parameter index"): triple (key, mask, offset).
NOTE 1. All the packets with IPv6 extension headers (but AH and ESP)
and all fragmented packets go to the best-effort traffic class.
NOTE 2. Two "port id"'s seems to be redundant, rfc2207 requires
only one "Generalized Port Identifier". So that for classic
ah, esp (and udp,tcp) both *pi should coincide or one of them
should be wildcard.
At first sight, this redundancy is just a waste of CPU
resources. But DPI and SPI add the possibility to assign different
priorities to GPIs. Look also at note 4 about tunnels below.
NOTE 3. One complication is the case of tunneled packets.
We implement it as following: if the first lookup
matches a special session with "tunnelhdr" value not zero,
flowid doesn't contain the true flow ID, but the tunnel ID (1...255).
In this case, we pull tunnelhdr bytes and restart lookup
with tunnel ID added to the list of keys. Simple and stupid 8)8)
It's enough for PIMREG and IPIP.
NOTE 4. Two GPIs make it possible to parse even GRE packets.
F.e. DPI can select ETH_P_IP (and necessary flags to make
tunnelhdr correct) in GRE protocol field and SPI matches
GRE key. Is it not nice? 8)8)
Well, as result, despite its simplicity, we get a pretty
powerful classification engine. */
#include <linux/config.h>
struct rsvp_head
{
u32 tmap[256/32];
u32 hgenerator;
u8 tgenerator;
struct rsvp_session *ht[256];
};
struct rsvp_session
{
struct rsvp_session *next;
u32 dst[RSVP_DST_LEN];
struct tc_rsvp_gpi dpi;
u8 protocol;
u8 tunnelid;
/* 16 (src,sport) hash slots, and one wildcard source slot */
struct rsvp_filter *ht[16+1];
};
struct rsvp_filter
{
struct rsvp_filter *next;
u32 src[RSVP_DST_LEN];
struct tc_rsvp_gpi spi;
u8 tunnelhdr;
struct tcf_result res;
struct tcf_exts exts;
u32 handle;
struct rsvp_session *sess;
};
static __inline__ unsigned hash_dst(u32 *dst, u8 protocol, u8 tunnelid)
{
unsigned h = dst[RSVP_DST_LEN-1];
h ^= h>>16;
h ^= h>>8;
return (h ^ protocol ^ tunnelid) & 0xFF;
}
static __inline__ unsigned hash_src(u32 *src)
{
unsigned h = src[RSVP_DST_LEN-1];
h ^= h>>16;
h ^= h>>8;
h ^= h>>4;
return h & 0xF;
}
static struct tcf_ext_map rsvp_ext_map = {
.police = TCA_RSVP_POLICE,
.action = TCA_RSVP_ACT
};
#define RSVP_APPLY_RESULT() \
{ \
int r = tcf_exts_exec(skb, &f->exts, res); \
if (r < 0) \
continue; \
else if (r > 0) \
return r; \
}
static int rsvp_classify(struct sk_buff *skb, struct tcf_proto *tp,
struct tcf_result *res)
{
struct rsvp_session **sht = ((struct rsvp_head*)tp->root)->ht;
struct rsvp_session *s;
struct rsvp_filter *f;
unsigned h1, h2;
u32 *dst, *src;
u8 protocol;
u8 tunnelid = 0;
u8 *xprt;
#if RSVP_DST_LEN == 4
struct ipv6hdr *nhptr = skb->nh.ipv6h;
#else
struct iphdr *nhptr = skb->nh.iph;
#endif
restart:
#if RSVP_DST_LEN == 4
src = &nhptr->saddr.s6_addr32[0];
dst = &nhptr->daddr.s6_addr32[0];
protocol = nhptr->nexthdr;
xprt = ((u8*)nhptr) + sizeof(struct ipv6hdr);
#else
src = &nhptr->saddr;
dst = &nhptr->daddr;
protocol = nhptr->protocol;
xprt = ((u8*)nhptr) + (nhptr->ihl<<2);
if (nhptr->frag_off&__constant_htons(IP_MF|IP_OFFSET))
return -1;
#endif
h1 = hash_dst(dst, protocol, tunnelid);
h2 = hash_src(src);
for (s = sht[h1]; s; s = s->next) {
if (dst[RSVP_DST_LEN-1] == s->dst[RSVP_DST_LEN-1] &&
protocol == s->protocol &&
!(s->dpi.mask & (*(u32*)(xprt+s->dpi.offset)^s->dpi.key))
#if RSVP_DST_LEN == 4
&& dst[0] == s->dst[0]
&& dst[1] == s->dst[1]
&& dst[2] == s->dst[2]
#endif
&& tunnelid == s->tunnelid) {
for (f = s->ht[h2]; f; f = f->next) {
if (src[RSVP_DST_LEN-1] == f->src[RSVP_DST_LEN-1] &&
!(f->spi.mask & (*(u32*)(xprt+f->spi.offset)^f->spi.key))
#if RSVP_DST_LEN == 4
&& src[0] == f->src[0]
&& src[1] == f->src[1]
&& src[2] == f->src[2]
#endif
) {
*res = f->res;
RSVP_APPLY_RESULT();
matched:
if (f->tunnelhdr == 0)
return 0;
tunnelid = f->res.classid;
nhptr = (void*)(xprt + f->tunnelhdr - sizeof(*nhptr));
goto restart;
}
}
/* And wildcard bucket... */
for (f = s->ht[16]; f; f = f->next) {
*res = f->res;
RSVP_APPLY_RESULT();
goto matched;
}
return -1;
}
}
return -1;
}
static unsigned long rsvp_get(struct tcf_proto *tp, u32 handle)
{
struct rsvp_session **sht = ((struct rsvp_head*)tp->root)->ht;
struct rsvp_session *s;
struct rsvp_filter *f;
unsigned h1 = handle&0xFF;
unsigned h2 = (handle>>8)&0xFF;
if (h2 > 16)
return 0;
for (s = sht[h1]; s; s = s->next) {
for (f = s->ht[h2]; f; f = f->next) {
if (f->handle == handle)
return (unsigned long)f;
}
}
return 0;
}
static void rsvp_put(struct tcf_proto *tp, unsigned long f)
{
}
static int rsvp_init(struct tcf_proto *tp)
{
struct rsvp_head *data;
data = kmalloc(sizeof(struct rsvp_head), GFP_KERNEL);
if (data) {
memset(data, 0, sizeof(struct rsvp_head));
tp->root = data;
return 0;
}
return -ENOBUFS;
}
static inline void
rsvp_delete_filter(struct tcf_proto *tp, struct rsvp_filter *f)
{
tcf_unbind_filter(tp, &f->res);
tcf_exts_destroy(tp, &f->exts);
kfree(f);
}
static void rsvp_destroy(struct tcf_proto *tp)
{
struct rsvp_head *data = xchg(&tp->root, NULL);
struct rsvp_session **sht;
int h1, h2;
if (data == NULL)
return;
sht = data->ht;
for (h1=0; h1<256; h1++) {
struct rsvp_session *s;
while ((s = sht[h1]) != NULL) {
sht[h1] = s->next;
for (h2=0; h2<=16; h2++) {
struct rsvp_filter *f;
while ((f = s->ht[h2]) != NULL) {
s->ht[h2] = f->next;
rsvp_delete_filter(tp, f);
}
}
kfree(s);
}
}
kfree(data);
}
static int rsvp_delete(struct tcf_proto *tp, unsigned long arg)
{
struct rsvp_filter **fp, *f = (struct rsvp_filter*)arg;
unsigned h = f->handle;
struct rsvp_session **sp;
struct rsvp_session *s = f->sess;
int i;
for (fp = &s->ht[(h>>8)&0xFF]; *fp; fp = &(*fp)->next) {
if (*fp == f) {
tcf_tree_lock(tp);
*fp = f->next;
tcf_tree_unlock(tp);
rsvp_delete_filter(tp, f);
/* Strip tree */
for (i=0; i<=16; i++)
if (s->ht[i])
return 0;
/* OK, session has no flows */
for (sp = &((struct rsvp_head*)tp->root)->ht[h&0xFF];
*sp; sp = &(*sp)->next) {
if (*sp == s) {
tcf_tree_lock(tp);
*sp = s->next;
tcf_tree_unlock(tp);
kfree(s);
return 0;
}
}
return 0;
}
}
return 0;
}
static unsigned gen_handle(struct tcf_proto *tp, unsigned salt)
{
struct rsvp_head *data = tp->root;
int i = 0xFFFF;
while (i-- > 0) {
u32 h;
if ((data->hgenerator += 0x10000) == 0)
data->hgenerator = 0x10000;
h = data->hgenerator|salt;
if (rsvp_get(tp, h) == 0)
return h;
}
return 0;
}
static int tunnel_bts(struct rsvp_head *data)
{
int n = data->tgenerator>>5;
u32 b = 1<<(data->tgenerator&0x1F);
if (data->tmap[n]&b)
return 0;
data->tmap[n] |= b;
return 1;
}
static void tunnel_recycle(struct rsvp_head *data)
{
struct rsvp_session **sht = data->ht;
u32 tmap[256/32];
int h1, h2;
memset(tmap, 0, sizeof(tmap));
for (h1=0; h1<256; h1++) {
struct rsvp_session *s;
for (s = sht[h1]; s; s = s->next) {
for (h2=0; h2<=16; h2++) {
struct rsvp_filter *f;
for (f = s->ht[h2]; f; f = f->next) {
if (f->tunnelhdr == 0)
continue;
data->tgenerator = f->res.classid;
tunnel_bts(data);
}
}
}
}
memcpy(data->tmap, tmap, sizeof(tmap));
}
static u32 gen_tunnel(struct rsvp_head *data)
{
int i, k;
for (k=0; k<2; k++) {
for (i=255; i>0; i--) {
if (++data->tgenerator == 0)
data->tgenerator = 1;
if (tunnel_bts(data))
return data->tgenerator;
}
tunnel_recycle(data);
}
return 0;
}
static int rsvp_change(struct tcf_proto *tp, unsigned long base,
u32 handle,
struct rtattr **tca,
unsigned long *arg)
{
struct rsvp_head *data = tp->root;
struct rsvp_filter *f, **fp;
struct rsvp_session *s, **sp;
struct tc_rsvp_pinfo *pinfo = NULL;
struct rtattr *opt = tca[TCA_OPTIONS-1];
struct rtattr *tb[TCA_RSVP_MAX];
struct tcf_exts e;
unsigned h1, h2;
u32 *dst;
int err;
if (opt == NULL)
return handle ? -EINVAL : 0;
if (rtattr_parse_nested(tb, TCA_RSVP_MAX, opt) < 0)
return -EINVAL;
err = tcf_exts_validate(tp, tb, tca[TCA_RATE-1], &e, &rsvp_ext_map);
if (err < 0)
return err;
if ((f = (struct rsvp_filter*)*arg) != NULL) {
/* Node exists: adjust only classid */
if (f->handle != handle && handle)
goto errout2;
if (tb[TCA_RSVP_CLASSID-1]) {
f->res.classid = *(u32*)RTA_DATA(tb[TCA_RSVP_CLASSID-1]);
tcf_bind_filter(tp, &f->res, base);
}
tcf_exts_change(tp, &f->exts, &e);
return 0;
}
/* Now more serious part... */
err = -EINVAL;
if (handle)
goto errout2;
if (tb[TCA_RSVP_DST-1] == NULL)
goto errout2;
err = -ENOBUFS;
f = kmalloc(sizeof(struct rsvp_filter), GFP_KERNEL);
if (f == NULL)
goto errout2;
memset(f, 0, sizeof(*f));
h2 = 16;
if (tb[TCA_RSVP_SRC-1]) {
err = -EINVAL;
if (RTA_PAYLOAD(tb[TCA_RSVP_SRC-1]) != sizeof(f->src))
goto errout;
memcpy(f->src, RTA_DATA(tb[TCA_RSVP_SRC-1]), sizeof(f->src));
h2 = hash_src(f->src);
}
if (tb[TCA_RSVP_PINFO-1]) {
err = -EINVAL;
if (RTA_PAYLOAD(tb[TCA_RSVP_PINFO-1]) < sizeof(struct tc_rsvp_pinfo))
goto errout;
pinfo = RTA_DATA(tb[TCA_RSVP_PINFO-1]);
f->spi = pinfo->spi;
f->tunnelhdr = pinfo->tunnelhdr;
}
if (tb[TCA_RSVP_CLASSID-1]) {
err = -EINVAL;
if (RTA_PAYLOAD(tb[TCA_RSVP_CLASSID-1]) != 4)
goto errout;
f->res.classid = *(u32*)RTA_DATA(tb[TCA_RSVP_CLASSID-1]);
}
err = -EINVAL;
if (RTA_PAYLOAD(tb[TCA_RSVP_DST-1]) != sizeof(f->src))
goto errout;
dst = RTA_DATA(tb[TCA_RSVP_DST-1]);
h1 = hash_dst(dst, pinfo ? pinfo->protocol : 0, pinfo ? pinfo->tunnelid : 0);
err = -ENOMEM;
if ((f->handle = gen_handle(tp, h1 | (h2<<8))) == 0)
goto errout;
if (f->tunnelhdr) {
err = -EINVAL;
if (f->res.classid > 255)
goto errout;
err = -ENOMEM;
if (f->res.classid == 0 &&
(f->res.classid = gen_tunnel(data)) == 0)
goto errout;
}
for (sp = &data->ht[h1]; (s=*sp) != NULL; sp = &s->next) {
if (dst[RSVP_DST_LEN-1] == s->dst[RSVP_DST_LEN-1] &&
pinfo && pinfo->protocol == s->protocol &&
memcmp(&pinfo->dpi, &s->dpi, sizeof(s->dpi)) == 0
#if RSVP_DST_LEN == 4
&& dst[0] == s->dst[0]
&& dst[1] == s->dst[1]
&& dst[2] == s->dst[2]
#endif
&& pinfo->tunnelid == s->tunnelid) {
insert:
/* OK, we found appropriate session */
fp = &s->ht[h2];
f->sess = s;
if (f->tunnelhdr == 0)
tcf_bind_filter(tp, &f->res, base);
tcf_exts_change(tp, &f->exts, &e);
for (fp = &s->ht[h2]; *fp; fp = &(*fp)->next)
if (((*fp)->spi.mask&f->spi.mask) != f->spi.mask)
break;
f->next = *fp;
wmb();
*fp = f;
*arg = (unsigned long)f;
return 0;
}
}
/* No session found. Create new one. */
err = -ENOBUFS;
s = kmalloc(sizeof(struct rsvp_session), GFP_KERNEL);
if (s == NULL)
goto errout;
memset(s, 0, sizeof(*s));
memcpy(s->dst, dst, sizeof(s->dst));
if (pinfo) {
s->dpi = pinfo->dpi;
s->protocol = pinfo->protocol;
s->tunnelid = pinfo->tunnelid;
}
for (sp = &data->ht[h1]; *sp; sp = &(*sp)->next) {
if (((*sp)->dpi.mask&s->dpi.mask) != s->dpi.mask)
break;
}
s->next = *sp;
wmb();
*sp = s;
goto insert;
errout:
if (f)
kfree(f);
errout2:
tcf_exts_destroy(tp, &e);
return err;
}
static void rsvp_walk(struct tcf_proto *tp, struct tcf_walker *arg)
{
struct rsvp_head *head = tp->root;
unsigned h, h1;
if (arg->stop)
return;
for (h = 0; h < 256; h++) {
struct rsvp_session *s;
for (s = head->ht[h]; s; s = s->next) {
for (h1 = 0; h1 <= 16; h1++) {
struct rsvp_filter *f;
for (f = s->ht[h1]; f; f = f->next) {
if (arg->count < arg->skip) {
arg->count++;
continue;
}
if (arg->fn(tp, (unsigned long)f, arg) < 0) {
arg->stop = 1;
return;
}
arg->count++;
}
}
}
}
}
static int rsvp_dump(struct tcf_proto *tp, unsigned long fh,
struct sk_buff *skb, struct tcmsg *t)
{
struct rsvp_filter *f = (struct rsvp_filter*)fh;
struct rsvp_session *s;
unsigned char *b = skb->tail;
struct rtattr *rta;
struct tc_rsvp_pinfo pinfo;
if (f == NULL)
return skb->len;
s = f->sess;
t->tcm_handle = f->handle;
rta = (struct rtattr*)b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
RTA_PUT(skb, TCA_RSVP_DST, sizeof(s->dst), &s->dst);
pinfo.dpi = s->dpi;
pinfo.spi = f->spi;
pinfo.protocol = s->protocol;
pinfo.tunnelid = s->tunnelid;
pinfo.tunnelhdr = f->tunnelhdr;
RTA_PUT(skb, TCA_RSVP_PINFO, sizeof(pinfo), &pinfo);
if (f->res.classid)
RTA_PUT(skb, TCA_RSVP_CLASSID, 4, &f->res.classid);
if (((f->handle>>8)&0xFF) != 16)
RTA_PUT(skb, TCA_RSVP_SRC, sizeof(f->src), f->src);
if (tcf_exts_dump(skb, &f->exts, &rsvp_ext_map) < 0)
goto rtattr_failure;
rta->rta_len = skb->tail - b;
if (tcf_exts_dump_stats(skb, &f->exts, &rsvp_ext_map) < 0)
goto rtattr_failure;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tcf_proto_ops RSVP_OPS = {
.next = NULL,
.kind = RSVP_ID,
.classify = rsvp_classify,
.init = rsvp_init,
.destroy = rsvp_destroy,
.get = rsvp_get,
.put = rsvp_put,
.change = rsvp_change,
.delete = rsvp_delete,
.walk = rsvp_walk,
.dump = rsvp_dump,
.owner = THIS_MODULE,
};
static int __init init_rsvp(void)
{
return register_tcf_proto_ops(&RSVP_OPS);
}
static void __exit exit_rsvp(void)
{
unregister_tcf_proto_ops(&RSVP_OPS);
}
module_init(init_rsvp)
module_exit(exit_rsvp)

44
net/sched/cls_rsvp6.c Normal file
View File

@@ -0,0 +1,44 @@
/*
* net/sched/cls_rsvp6.c Special RSVP packet classifier for IPv6.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <linux/ipv6.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>
#define RSVP_DST_LEN 4
#define RSVP_ID "rsvp6"
#define RSVP_OPS cls_rsvp6_ops
#include "cls_rsvp.h"
MODULE_LICENSE("GPL");

537
net/sched/cls_tcindex.c Normal file
View File

@@ -0,0 +1,537 @@
/*
* net/sched/cls_tcindex.c Packet classifier for skb->tc_index
*
* Written 1998,1999 by Werner Almesberger, EPFL ICA
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/skbuff.h>
#include <linux/errno.h>
#include <linux/netdevice.h>
#include <net/ip.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>
#include <net/route.h>
/*
* Not quite sure if we need all the xchgs Alexey uses when accessing things.
* Can always add them later ... :)
*/
/*
* Passing parameters to the root seems to be done more awkwardly than really
* necessary. At least, u32 doesn't seem to use such dirty hacks. To be
* verified. FIXME.
*/
#define PERFECT_HASH_THRESHOLD 64 /* use perfect hash if not bigger */
#define DEFAULT_HASH_SIZE 64 /* optimized for diffserv */
#if 1 /* control */
#define DPRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define DPRINTK(format,args...)
#endif
#if 0 /* data */
#define D2PRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define D2PRINTK(format,args...)
#endif
#define PRIV(tp) ((struct tcindex_data *) (tp)->root)
struct tcindex_filter_result {
struct tcf_exts exts;
struct tcf_result res;
};
struct tcindex_filter {
u16 key;
struct tcindex_filter_result result;
struct tcindex_filter *next;
};
struct tcindex_data {
struct tcindex_filter_result *perfect; /* perfect hash; NULL if none */
struct tcindex_filter **h; /* imperfect hash; only used if !perfect;
NULL if unused */
u16 mask; /* AND key with mask */
int shift; /* shift ANDed key to the right */
int hash; /* hash table size; 0 if undefined */
int alloc_hash; /* allocated size */
int fall_through; /* 0: only classify if explicit match */
};
static struct tcf_ext_map tcindex_ext_map = {
.police = TCA_TCINDEX_POLICE,
.action = TCA_TCINDEX_ACT
};
static inline int
tcindex_filter_is_set(struct tcindex_filter_result *r)
{
return tcf_exts_is_predicative(&r->exts) || r->res.classid;
}
static struct tcindex_filter_result *
tcindex_lookup(struct tcindex_data *p, u16 key)
{
struct tcindex_filter *f;
if (p->perfect)
return tcindex_filter_is_set(p->perfect + key) ?
p->perfect + key : NULL;
else if (p->h) {
for (f = p->h[key % p->hash]; f; f = f->next)
if (f->key == key)
return &f->result;
}
return NULL;
}
static int tcindex_classify(struct sk_buff *skb, struct tcf_proto *tp,
struct tcf_result *res)
{
struct tcindex_data *p = PRIV(tp);
struct tcindex_filter_result *f;
int key = (skb->tc_index & p->mask) >> p->shift;
D2PRINTK("tcindex_classify(skb %p,tp %p,res %p),p %p\n",skb,tp,res,p);
f = tcindex_lookup(p, key);
if (!f) {
if (!p->fall_through)
return -1;
res->classid = TC_H_MAKE(TC_H_MAJ(tp->q->handle), key);
res->class = 0;
D2PRINTK("alg 0x%x\n",res->classid);
return 0;
}
*res = f->res;
D2PRINTK("map 0x%x\n",res->classid);
return tcf_exts_exec(skb, &f->exts, res);
}
static unsigned long tcindex_get(struct tcf_proto *tp, u32 handle)
{
struct tcindex_data *p = PRIV(tp);
struct tcindex_filter_result *r;
DPRINTK("tcindex_get(tp %p,handle 0x%08x)\n",tp,handle);
if (p->perfect && handle >= p->alloc_hash)
return 0;
r = tcindex_lookup(p, handle);
return r && tcindex_filter_is_set(r) ? (unsigned long) r : 0UL;
}
static void tcindex_put(struct tcf_proto *tp, unsigned long f)
{
DPRINTK("tcindex_put(tp %p,f 0x%lx)\n",tp,f);
}
static int tcindex_init(struct tcf_proto *tp)
{
struct tcindex_data *p;
DPRINTK("tcindex_init(tp %p)\n",tp);
p = kmalloc(sizeof(struct tcindex_data),GFP_KERNEL);
if (!p)
return -ENOMEM;
memset(p, 0, sizeof(*p));
p->mask = 0xffff;
p->hash = DEFAULT_HASH_SIZE;
p->fall_through = 1;
tp->root = p;
return 0;
}
static int
__tcindex_delete(struct tcf_proto *tp, unsigned long arg, int lock)
{
struct tcindex_data *p = PRIV(tp);
struct tcindex_filter_result *r = (struct tcindex_filter_result *) arg;
struct tcindex_filter *f = NULL;
DPRINTK("tcindex_delete(tp %p,arg 0x%lx),p %p,f %p\n",tp,arg,p,f);
if (p->perfect) {
if (!r->res.class)
return -ENOENT;
} else {
int i;
struct tcindex_filter **walk = NULL;
for (i = 0; i < p->hash; i++)
for (walk = p->h+i; *walk; walk = &(*walk)->next)
if (&(*walk)->result == r)
goto found;
return -ENOENT;
found:
f = *walk;
if (lock)
tcf_tree_lock(tp);
*walk = f->next;
if (lock)
tcf_tree_unlock(tp);
}
tcf_unbind_filter(tp, &r->res);
tcf_exts_destroy(tp, &r->exts);
if (f)
kfree(f);
return 0;
}
static int tcindex_delete(struct tcf_proto *tp, unsigned long arg)
{
return __tcindex_delete(tp, arg, 1);
}
static inline int
valid_perfect_hash(struct tcindex_data *p)
{
return p->hash > (p->mask >> p->shift);
}
static int
tcindex_set_parms(struct tcf_proto *tp, unsigned long base, u32 handle,
struct tcindex_data *p, struct tcindex_filter_result *r,
struct rtattr **tb, struct rtattr *est)
{
int err, balloc = 0;
struct tcindex_filter_result new_filter_result, *old_r = r;
struct tcindex_filter_result cr;
struct tcindex_data cp;
struct tcindex_filter *f = NULL; /* make gcc behave */
struct tcf_exts e;
err = tcf_exts_validate(tp, tb, est, &e, &tcindex_ext_map);
if (err < 0)
return err;
memcpy(&cp, p, sizeof(cp));
memset(&new_filter_result, 0, sizeof(new_filter_result));
if (old_r)
memcpy(&cr, r, sizeof(cr));
else
memset(&cr, 0, sizeof(cr));
err = -EINVAL;
if (tb[TCA_TCINDEX_HASH-1]) {
if (RTA_PAYLOAD(tb[TCA_TCINDEX_HASH-1]) < sizeof(u32))
goto errout;
cp.hash = *(u32 *) RTA_DATA(tb[TCA_TCINDEX_HASH-1]);
}
if (tb[TCA_TCINDEX_MASK-1]) {
if (RTA_PAYLOAD(tb[TCA_TCINDEX_MASK-1]) < sizeof(u16))
goto errout;
cp.mask = *(u16 *) RTA_DATA(tb[TCA_TCINDEX_MASK-1]);
}
if (tb[TCA_TCINDEX_SHIFT-1]) {
if (RTA_PAYLOAD(tb[TCA_TCINDEX_SHIFT-1]) < sizeof(u16))
goto errout;
cp.shift = *(u16 *) RTA_DATA(tb[TCA_TCINDEX_SHIFT-1]);
}
err = -EBUSY;
/* Hash already allocated, make sure that we still meet the
* requirements for the allocated hash.
*/
if (cp.perfect) {
if (!valid_perfect_hash(&cp) ||
cp.hash > cp.alloc_hash)
goto errout;
} else if (cp.h && cp.hash != cp.alloc_hash)
goto errout;
err = -EINVAL;
if (tb[TCA_TCINDEX_FALL_THROUGH-1]) {
if (RTA_PAYLOAD(tb[TCA_TCINDEX_FALL_THROUGH-1]) < sizeof(u32))
goto errout;
cp.fall_through =
*(u32 *) RTA_DATA(tb[TCA_TCINDEX_FALL_THROUGH-1]);
}
if (!cp.hash) {
/* Hash not specified, use perfect hash if the upper limit
* of the hashing index is below the threshold.
*/
if ((cp.mask >> cp.shift) < PERFECT_HASH_THRESHOLD)
cp.hash = (cp.mask >> cp.shift)+1;
else
cp.hash = DEFAULT_HASH_SIZE;
}
if (!cp.perfect && !cp.h)
cp.alloc_hash = cp.hash;
/* Note: this could be as restrictive as if (handle & ~(mask >> shift))
* but then, we'd fail handles that may become valid after some future
* mask change. While this is extremely unlikely to ever matter,
* the check below is safer (and also more backwards-compatible).
*/
if (cp.perfect || valid_perfect_hash(&cp))
if (handle >= cp.alloc_hash)
goto errout;
err = -ENOMEM;
if (!cp.perfect && !cp.h) {
if (valid_perfect_hash(&cp)) {
cp.perfect = kmalloc(cp.hash * sizeof(*r), GFP_KERNEL);
if (!cp.perfect)
goto errout;
memset(cp.perfect, 0, cp.hash * sizeof(*r));
balloc = 1;
} else {
cp.h = kmalloc(cp.hash * sizeof(f), GFP_KERNEL);
if (!cp.h)
goto errout;
memset(cp.h, 0, cp.hash * sizeof(f));
balloc = 2;
}
}
if (cp.perfect)
r = cp.perfect + handle;
else
r = tcindex_lookup(&cp, handle) ? : &new_filter_result;
if (r == &new_filter_result) {
f = kmalloc(sizeof(*f), GFP_KERNEL);
if (!f)
goto errout_alloc;
memset(f, 0, sizeof(*f));
}
if (tb[TCA_TCINDEX_CLASSID-1]) {
cr.res.classid = *(u32 *) RTA_DATA(tb[TCA_TCINDEX_CLASSID-1]);
tcf_bind_filter(tp, &cr.res, base);
}
tcf_exts_change(tp, &cr.exts, &e);
tcf_tree_lock(tp);
if (old_r && old_r != r)
memset(old_r, 0, sizeof(*old_r));
memcpy(p, &cp, sizeof(cp));
memcpy(r, &cr, sizeof(cr));
if (r == &new_filter_result) {
struct tcindex_filter **fp;
f->key = handle;
f->result = new_filter_result;
f->next = NULL;
for (fp = p->h+(handle % p->hash); *fp; fp = &(*fp)->next)
/* nothing */;
*fp = f;
}
tcf_tree_unlock(tp);
return 0;
errout_alloc:
if (balloc == 1)
kfree(cp.perfect);
else if (balloc == 2)
kfree(cp.h);
errout:
tcf_exts_destroy(tp, &e);
return err;
}
static int
tcindex_change(struct tcf_proto *tp, unsigned long base, u32 handle,
struct rtattr **tca, unsigned long *arg)
{
struct rtattr *opt = tca[TCA_OPTIONS-1];
struct rtattr *tb[TCA_TCINDEX_MAX];
struct tcindex_data *p = PRIV(tp);
struct tcindex_filter_result *r = (struct tcindex_filter_result *) *arg;
DPRINTK("tcindex_change(tp %p,handle 0x%08x,tca %p,arg %p),opt %p,"
"p %p,r %p,*arg 0x%lx\n",
tp, handle, tca, arg, opt, p, r, arg ? *arg : 0L);
if (!opt)
return 0;
if (rtattr_parse_nested(tb, TCA_TCINDEX_MAX, opt) < 0)
return -EINVAL;
return tcindex_set_parms(tp, base, handle, p, r, tb, tca[TCA_RATE-1]);
}
static void tcindex_walk(struct tcf_proto *tp, struct tcf_walker *walker)
{
struct tcindex_data *p = PRIV(tp);
struct tcindex_filter *f,*next;
int i;
DPRINTK("tcindex_walk(tp %p,walker %p),p %p\n",tp,walker,p);
if (p->perfect) {
for (i = 0; i < p->hash; i++) {
if (!p->perfect[i].res.class)
continue;
if (walker->count >= walker->skip) {
if (walker->fn(tp,
(unsigned long) (p->perfect+i), walker)
< 0) {
walker->stop = 1;
return;
}
}
walker->count++;
}
}
if (!p->h)
return;
for (i = 0; i < p->hash; i++) {
for (f = p->h[i]; f; f = next) {
next = f->next;
if (walker->count >= walker->skip) {
if (walker->fn(tp,(unsigned long) &f->result,
walker) < 0) {
walker->stop = 1;
return;
}
}
walker->count++;
}
}
}
static int tcindex_destroy_element(struct tcf_proto *tp,
unsigned long arg, struct tcf_walker *walker)
{
return __tcindex_delete(tp, arg, 0);
}
static void tcindex_destroy(struct tcf_proto *tp)
{
struct tcindex_data *p = PRIV(tp);
struct tcf_walker walker;
DPRINTK("tcindex_destroy(tp %p),p %p\n",tp,p);
walker.count = 0;
walker.skip = 0;
walker.fn = &tcindex_destroy_element;
tcindex_walk(tp,&walker);
if (p->perfect)
kfree(p->perfect);
if (p->h)
kfree(p->h);
kfree(p);
tp->root = NULL;
}
static int tcindex_dump(struct tcf_proto *tp, unsigned long fh,
struct sk_buff *skb, struct tcmsg *t)
{
struct tcindex_data *p = PRIV(tp);
struct tcindex_filter_result *r = (struct tcindex_filter_result *) fh;
unsigned char *b = skb->tail;
struct rtattr *rta;
DPRINTK("tcindex_dump(tp %p,fh 0x%lx,skb %p,t %p),p %p,r %p,b %p\n",
tp,fh,skb,t,p,r,b);
DPRINTK("p->perfect %p p->h %p\n",p->perfect,p->h);
rta = (struct rtattr *) b;
RTA_PUT(skb,TCA_OPTIONS,0,NULL);
if (!fh) {
t->tcm_handle = ~0; /* whatever ... */
RTA_PUT(skb,TCA_TCINDEX_HASH,sizeof(p->hash),&p->hash);
RTA_PUT(skb,TCA_TCINDEX_MASK,sizeof(p->mask),&p->mask);
RTA_PUT(skb,TCA_TCINDEX_SHIFT,sizeof(p->shift),&p->shift);
RTA_PUT(skb,TCA_TCINDEX_FALL_THROUGH,sizeof(p->fall_through),
&p->fall_through);
rta->rta_len = skb->tail-b;
} else {
if (p->perfect) {
t->tcm_handle = r-p->perfect;
} else {
struct tcindex_filter *f;
int i;
t->tcm_handle = 0;
for (i = 0; !t->tcm_handle && i < p->hash; i++) {
for (f = p->h[i]; !t->tcm_handle && f;
f = f->next) {
if (&f->result == r)
t->tcm_handle = f->key;
}
}
}
DPRINTK("handle = %d\n",t->tcm_handle);
if (r->res.class)
RTA_PUT(skb, TCA_TCINDEX_CLASSID, 4, &r->res.classid);
if (tcf_exts_dump(skb, &r->exts, &tcindex_ext_map) < 0)
goto rtattr_failure;
rta->rta_len = skb->tail-b;
if (tcf_exts_dump_stats(skb, &r->exts, &tcindex_ext_map) < 0)
goto rtattr_failure;
}
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tcf_proto_ops cls_tcindex_ops = {
.next = NULL,
.kind = "tcindex",
.classify = tcindex_classify,
.init = tcindex_init,
.destroy = tcindex_destroy,
.get = tcindex_get,
.put = tcindex_put,
.change = tcindex_change,
.delete = tcindex_delete,
.walk = tcindex_walk,
.dump = tcindex_dump,
.owner = THIS_MODULE,
};
static int __init init_tcindex(void)
{
return register_tcf_proto_ops(&cls_tcindex_ops);
}
static void __exit exit_tcindex(void)
{
unregister_tcf_proto_ops(&cls_tcindex_ops);
}
module_init(init_tcindex)
module_exit(exit_tcindex)
MODULE_LICENSE("GPL");

828
net/sched/cls_u32.c Normal file
View File

@@ -0,0 +1,828 @@
/*
* net/sched/cls_u32.c Ugly (or Universal) 32bit key Packet Classifier.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
* The filters are packed to hash tables of key nodes
* with a set of 32bit key/mask pairs at every node.
* Nodes reference next level hash tables etc.
*
* This scheme is the best universal classifier I managed to
* invent; it is not super-fast, but it is not slow (provided you
* program it correctly), and general enough. And its relative
* speed grows as the number of rules becomes larger.
*
* It seems that it represents the best middle point between
* speed and manageability both by human and by machine.
*
* It is especially useful for link sharing combined with QoS;
* pure RSVP doesn't need such a general approach and can use
* much simpler (and faster) schemes, sort of cls_rsvp.c.
*
* JHS: We should remove the CONFIG_NET_CLS_IND from here
* eventually when the meta match extension is made available
*
* nfmark match added by Catalin(ux aka Dino) BOIE <catab at umbrella.ro>
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <linux/rtnetlink.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>
struct tc_u_knode
{
struct tc_u_knode *next;
u32 handle;
struct tc_u_hnode *ht_up;
struct tcf_exts exts;
#ifdef CONFIG_NET_CLS_IND
char indev[IFNAMSIZ];
#endif
u8 fshift;
struct tcf_result res;
struct tc_u_hnode *ht_down;
#ifdef CONFIG_CLS_U32_PERF
struct tc_u32_pcnt *pf;
#endif
#ifdef CONFIG_CLS_U32_MARK
struct tc_u32_mark mark;
#endif
struct tc_u32_sel sel;
};
struct tc_u_hnode
{
struct tc_u_hnode *next;
u32 handle;
u32 prio;
struct tc_u_common *tp_c;
int refcnt;
unsigned divisor;
struct tc_u_knode *ht[1];
};
struct tc_u_common
{
struct tc_u_common *next;
struct tc_u_hnode *hlist;
struct Qdisc *q;
int refcnt;
u32 hgenerator;
};
static struct tcf_ext_map u32_ext_map = {
.action = TCA_U32_ACT,
.police = TCA_U32_POLICE
};
static struct tc_u_common *u32_list;
static __inline__ unsigned u32_hash_fold(u32 key, struct tc_u32_sel *sel, u8 fshift)
{
unsigned h = (key & sel->hmask)>>fshift;
return h;
}
static int u32_classify(struct sk_buff *skb, struct tcf_proto *tp, struct tcf_result *res)
{
struct {
struct tc_u_knode *knode;
u8 *ptr;
} stack[TC_U32_MAXDEPTH];
struct tc_u_hnode *ht = (struct tc_u_hnode*)tp->root;
u8 *ptr = skb->nh.raw;
struct tc_u_knode *n;
int sdepth = 0;
int off2 = 0;
int sel = 0;
#ifdef CONFIG_CLS_U32_PERF
int j;
#endif
int i, r;
next_ht:
n = ht->ht[sel];
next_knode:
if (n) {
struct tc_u32_key *key = n->sel.keys;
#ifdef CONFIG_CLS_U32_PERF
n->pf->rcnt +=1;
j = 0;
#endif
#ifdef CONFIG_CLS_U32_MARK
if ((skb->nfmark & n->mark.mask) != n->mark.val) {
n = n->next;
goto next_knode;
} else {
n->mark.success++;
}
#endif
for (i = n->sel.nkeys; i>0; i--, key++) {
if ((*(u32*)(ptr+key->off+(off2&key->offmask))^key->val)&key->mask) {
n = n->next;
goto next_knode;
}
#ifdef CONFIG_CLS_U32_PERF
n->pf->kcnts[j] +=1;
j++;
#endif
}
if (n->ht_down == NULL) {
check_terminal:
if (n->sel.flags&TC_U32_TERMINAL) {
*res = n->res;
#ifdef CONFIG_NET_CLS_IND
if (!tcf_match_indev(skb, n->indev)) {
n = n->next;
goto next_knode;
}
#endif
#ifdef CONFIG_CLS_U32_PERF
n->pf->rhit +=1;
#endif
r = tcf_exts_exec(skb, &n->exts, res);
if (r < 0) {
n = n->next;
goto next_knode;
}
return r;
}
n = n->next;
goto next_knode;
}
/* PUSH */
if (sdepth >= TC_U32_MAXDEPTH)
goto deadloop;
stack[sdepth].knode = n;
stack[sdepth].ptr = ptr;
sdepth++;
ht = n->ht_down;
sel = 0;
if (ht->divisor)
sel = ht->divisor&u32_hash_fold(*(u32*)(ptr+n->sel.hoff), &n->sel,n->fshift);
if (!(n->sel.flags&(TC_U32_VAROFFSET|TC_U32_OFFSET|TC_U32_EAT)))
goto next_ht;
if (n->sel.flags&(TC_U32_OFFSET|TC_U32_VAROFFSET)) {
off2 = n->sel.off + 3;
if (n->sel.flags&TC_U32_VAROFFSET)
off2 += ntohs(n->sel.offmask & *(u16*)(ptr+n->sel.offoff)) >>n->sel.offshift;
off2 &= ~3;
}
if (n->sel.flags&TC_U32_EAT) {
ptr += off2;
off2 = 0;
}
if (ptr < skb->tail)
goto next_ht;
}
/* POP */
if (sdepth--) {
n = stack[sdepth].knode;
ht = n->ht_up;
ptr = stack[sdepth].ptr;
goto check_terminal;
}
return -1;
deadloop:
if (net_ratelimit())
printk("cls_u32: dead loop\n");
return -1;
}
static __inline__ struct tc_u_hnode *
u32_lookup_ht(struct tc_u_common *tp_c, u32 handle)
{
struct tc_u_hnode *ht;
for (ht = tp_c->hlist; ht; ht = ht->next)
if (ht->handle == handle)
break;
return ht;
}
static __inline__ struct tc_u_knode *
u32_lookup_key(struct tc_u_hnode *ht, u32 handle)
{
unsigned sel;
struct tc_u_knode *n = NULL;
sel = TC_U32_HASH(handle);
if (sel > ht->divisor)
goto out;
for (n = ht->ht[sel]; n; n = n->next)
if (n->handle == handle)
break;
out:
return n;
}
static unsigned long u32_get(struct tcf_proto *tp, u32 handle)
{
struct tc_u_hnode *ht;
struct tc_u_common *tp_c = tp->data;
if (TC_U32_HTID(handle) == TC_U32_ROOT)
ht = tp->root;
else
ht = u32_lookup_ht(tp_c, TC_U32_HTID(handle));
if (!ht)
return 0;
if (TC_U32_KEY(handle) == 0)
return (unsigned long)ht;
return (unsigned long)u32_lookup_key(ht, handle);
}
static void u32_put(struct tcf_proto *tp, unsigned long f)
{
}
static u32 gen_new_htid(struct tc_u_common *tp_c)
{
int i = 0x800;
do {
if (++tp_c->hgenerator == 0x7FF)
tp_c->hgenerator = 1;
} while (--i>0 && u32_lookup_ht(tp_c, (tp_c->hgenerator|0x800)<<20));
return i > 0 ? (tp_c->hgenerator|0x800)<<20 : 0;
}
static int u32_init(struct tcf_proto *tp)
{
struct tc_u_hnode *root_ht;
struct tc_u_common *tp_c;
for (tp_c = u32_list; tp_c; tp_c = tp_c->next)
if (tp_c->q == tp->q)
break;
root_ht = kmalloc(sizeof(*root_ht), GFP_KERNEL);
if (root_ht == NULL)
return -ENOBUFS;
memset(root_ht, 0, sizeof(*root_ht));
root_ht->divisor = 0;
root_ht->refcnt++;
root_ht->handle = tp_c ? gen_new_htid(tp_c) : 0x80000000;
root_ht->prio = tp->prio;
if (tp_c == NULL) {
tp_c = kmalloc(sizeof(*tp_c), GFP_KERNEL);
if (tp_c == NULL) {
kfree(root_ht);
return -ENOBUFS;
}
memset(tp_c, 0, sizeof(*tp_c));
tp_c->q = tp->q;
tp_c->next = u32_list;
u32_list = tp_c;
}
tp_c->refcnt++;
root_ht->next = tp_c->hlist;
tp_c->hlist = root_ht;
root_ht->tp_c = tp_c;
tp->root = root_ht;
tp->data = tp_c;
return 0;
}
static int u32_destroy_key(struct tcf_proto *tp, struct tc_u_knode *n)
{
tcf_unbind_filter(tp, &n->res);
tcf_exts_destroy(tp, &n->exts);
if (n->ht_down)
n->ht_down->refcnt--;
#ifdef CONFIG_CLS_U32_PERF
if (n && (NULL != n->pf))
kfree(n->pf);
#endif
kfree(n);
return 0;
}
static int u32_delete_key(struct tcf_proto *tp, struct tc_u_knode* key)
{
struct tc_u_knode **kp;
struct tc_u_hnode *ht = key->ht_up;
if (ht) {
for (kp = &ht->ht[TC_U32_HASH(key->handle)]; *kp; kp = &(*kp)->next) {
if (*kp == key) {
tcf_tree_lock(tp);
*kp = key->next;
tcf_tree_unlock(tp);
u32_destroy_key(tp, key);
return 0;
}
}
}
BUG_TRAP(0);
return 0;
}
static void u32_clear_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht)
{
struct tc_u_knode *n;
unsigned h;
for (h=0; h<=ht->divisor; h++) {
while ((n = ht->ht[h]) != NULL) {
ht->ht[h] = n->next;
u32_destroy_key(tp, n);
}
}
}
static int u32_destroy_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht)
{
struct tc_u_common *tp_c = tp->data;
struct tc_u_hnode **hn;
BUG_TRAP(!ht->refcnt);
u32_clear_hnode(tp, ht);
for (hn = &tp_c->hlist; *hn; hn = &(*hn)->next) {
if (*hn == ht) {
*hn = ht->next;
kfree(ht);
return 0;
}
}
BUG_TRAP(0);
return -ENOENT;
}
static void u32_destroy(struct tcf_proto *tp)
{
struct tc_u_common *tp_c = tp->data;
struct tc_u_hnode *root_ht = xchg(&tp->root, NULL);
BUG_TRAP(root_ht != NULL);
if (root_ht && --root_ht->refcnt == 0)
u32_destroy_hnode(tp, root_ht);
if (--tp_c->refcnt == 0) {
struct tc_u_hnode *ht;
struct tc_u_common **tp_cp;
for (tp_cp = &u32_list; *tp_cp; tp_cp = &(*tp_cp)->next) {
if (*tp_cp == tp_c) {
*tp_cp = tp_c->next;
break;
}
}
for (ht=tp_c->hlist; ht; ht = ht->next)
u32_clear_hnode(tp, ht);
while ((ht = tp_c->hlist) != NULL) {
tp_c->hlist = ht->next;
BUG_TRAP(ht->refcnt == 0);
kfree(ht);
};
kfree(tp_c);
}
tp->data = NULL;
}
static int u32_delete(struct tcf_proto *tp, unsigned long arg)
{
struct tc_u_hnode *ht = (struct tc_u_hnode*)arg;
if (ht == NULL)
return 0;
if (TC_U32_KEY(ht->handle))
return u32_delete_key(tp, (struct tc_u_knode*)ht);
if (tp->root == ht)
return -EINVAL;
if (--ht->refcnt == 0)
u32_destroy_hnode(tp, ht);
return 0;
}
static u32 gen_new_kid(struct tc_u_hnode *ht, u32 handle)
{
struct tc_u_knode *n;
unsigned i = 0x7FF;
for (n=ht->ht[TC_U32_HASH(handle)]; n; n = n->next)
if (i < TC_U32_NODE(n->handle))
i = TC_U32_NODE(n->handle);
i++;
return handle|(i>0xFFF ? 0xFFF : i);
}
static int u32_set_parms(struct tcf_proto *tp, unsigned long base,
struct tc_u_hnode *ht,
struct tc_u_knode *n, struct rtattr **tb,
struct rtattr *est)
{
int err;
struct tcf_exts e;
err = tcf_exts_validate(tp, tb, est, &e, &u32_ext_map);
if (err < 0)
return err;
err = -EINVAL;
if (tb[TCA_U32_LINK-1]) {
u32 handle = *(u32*)RTA_DATA(tb[TCA_U32_LINK-1]);
struct tc_u_hnode *ht_down = NULL;
if (TC_U32_KEY(handle))
goto errout;
if (handle) {
ht_down = u32_lookup_ht(ht->tp_c, handle);
if (ht_down == NULL)
goto errout;
ht_down->refcnt++;
}
tcf_tree_lock(tp);
ht_down = xchg(&n->ht_down, ht_down);
tcf_tree_unlock(tp);
if (ht_down)
ht_down->refcnt--;
}
if (tb[TCA_U32_CLASSID-1]) {
n->res.classid = *(u32*)RTA_DATA(tb[TCA_U32_CLASSID-1]);
tcf_bind_filter(tp, &n->res, base);
}
#ifdef CONFIG_NET_CLS_IND
if (tb[TCA_U32_INDEV-1]) {
int err = tcf_change_indev(tp, n->indev, tb[TCA_U32_INDEV-1]);
if (err < 0)
goto errout;
}
#endif
tcf_exts_change(tp, &n->exts, &e);
return 0;
errout:
tcf_exts_destroy(tp, &e);
return err;
}
static int u32_change(struct tcf_proto *tp, unsigned long base, u32 handle,
struct rtattr **tca,
unsigned long *arg)
{
struct tc_u_common *tp_c = tp->data;
struct tc_u_hnode *ht;
struct tc_u_knode *n;
struct tc_u32_sel *s;
struct rtattr *opt = tca[TCA_OPTIONS-1];
struct rtattr *tb[TCA_U32_MAX];
u32 htid;
int err;
if (opt == NULL)
return handle ? -EINVAL : 0;
if (rtattr_parse_nested(tb, TCA_U32_MAX, opt) < 0)
return -EINVAL;
if ((n = (struct tc_u_knode*)*arg) != NULL) {
if (TC_U32_KEY(n->handle) == 0)
return -EINVAL;
return u32_set_parms(tp, base, n->ht_up, n, tb, tca[TCA_RATE-1]);
}
if (tb[TCA_U32_DIVISOR-1]) {
unsigned divisor = *(unsigned*)RTA_DATA(tb[TCA_U32_DIVISOR-1]);
if (--divisor > 0x100)
return -EINVAL;
if (TC_U32_KEY(handle))
return -EINVAL;
if (handle == 0) {
handle = gen_new_htid(tp->data);
if (handle == 0)
return -ENOMEM;
}
ht = kmalloc(sizeof(*ht) + divisor*sizeof(void*), GFP_KERNEL);
if (ht == NULL)
return -ENOBUFS;
memset(ht, 0, sizeof(*ht) + divisor*sizeof(void*));
ht->tp_c = tp_c;
ht->refcnt = 0;
ht->divisor = divisor;
ht->handle = handle;
ht->prio = tp->prio;
ht->next = tp_c->hlist;
tp_c->hlist = ht;
*arg = (unsigned long)ht;
return 0;
}
if (tb[TCA_U32_HASH-1]) {
htid = *(unsigned*)RTA_DATA(tb[TCA_U32_HASH-1]);
if (TC_U32_HTID(htid) == TC_U32_ROOT) {
ht = tp->root;
htid = ht->handle;
} else {
ht = u32_lookup_ht(tp->data, TC_U32_HTID(htid));
if (ht == NULL)
return -EINVAL;
}
} else {
ht = tp->root;
htid = ht->handle;
}
if (ht->divisor < TC_U32_HASH(htid))
return -EINVAL;
if (handle) {
if (TC_U32_HTID(handle) && TC_U32_HTID(handle^htid))
return -EINVAL;
handle = htid | TC_U32_NODE(handle);
} else
handle = gen_new_kid(ht, htid);
if (tb[TCA_U32_SEL-1] == 0 ||
RTA_PAYLOAD(tb[TCA_U32_SEL-1]) < sizeof(struct tc_u32_sel))
return -EINVAL;
s = RTA_DATA(tb[TCA_U32_SEL-1]);
n = kmalloc(sizeof(*n) + s->nkeys*sizeof(struct tc_u32_key), GFP_KERNEL);
if (n == NULL)
return -ENOBUFS;
memset(n, 0, sizeof(*n) + s->nkeys*sizeof(struct tc_u32_key));
#ifdef CONFIG_CLS_U32_PERF
n->pf = kmalloc(sizeof(struct tc_u32_pcnt) + s->nkeys*sizeof(u64), GFP_KERNEL);
if (n->pf == NULL) {
kfree(n);
return -ENOBUFS;
}
memset(n->pf, 0, sizeof(struct tc_u32_pcnt) + s->nkeys*sizeof(u64));
#endif
memcpy(&n->sel, s, sizeof(*s) + s->nkeys*sizeof(struct tc_u32_key));
n->ht_up = ht;
n->handle = handle;
{
u8 i = 0;
u32 mask = s->hmask;
if (mask) {
while (!(mask & 1)) {
i++;
mask>>=1;
}
}
n->fshift = i;
}
#ifdef CONFIG_CLS_U32_MARK
if (tb[TCA_U32_MARK-1]) {
struct tc_u32_mark *mark;
if (RTA_PAYLOAD(tb[TCA_U32_MARK-1]) < sizeof(struct tc_u32_mark)) {
#ifdef CONFIG_CLS_U32_PERF
kfree(n->pf);
#endif
kfree(n);
return -EINVAL;
}
mark = RTA_DATA(tb[TCA_U32_MARK-1]);
memcpy(&n->mark, mark, sizeof(struct tc_u32_mark));
n->mark.success = 0;
}
#endif
err = u32_set_parms(tp, base, ht, n, tb, tca[TCA_RATE-1]);
if (err == 0) {
struct tc_u_knode **ins;
for (ins = &ht->ht[TC_U32_HASH(handle)]; *ins; ins = &(*ins)->next)
if (TC_U32_NODE(handle) < TC_U32_NODE((*ins)->handle))
break;
n->next = *ins;
wmb();
*ins = n;
*arg = (unsigned long)n;
return 0;
}
#ifdef CONFIG_CLS_U32_PERF
if (n && (NULL != n->pf))
kfree(n->pf);
#endif
kfree(n);
return err;
}
static void u32_walk(struct tcf_proto *tp, struct tcf_walker *arg)
{
struct tc_u_common *tp_c = tp->data;
struct tc_u_hnode *ht;
struct tc_u_knode *n;
unsigned h;
if (arg->stop)
return;
for (ht = tp_c->hlist; ht; ht = ht->next) {
if (ht->prio != tp->prio)
continue;
if (arg->count >= arg->skip) {
if (arg->fn(tp, (unsigned long)ht, arg) < 0) {
arg->stop = 1;
return;
}
}
arg->count++;
for (h = 0; h <= ht->divisor; h++) {
for (n = ht->ht[h]; n; n = n->next) {
if (arg->count < arg->skip) {
arg->count++;
continue;
}
if (arg->fn(tp, (unsigned long)n, arg) < 0) {
arg->stop = 1;
return;
}
arg->count++;
}
}
}
}
static int u32_dump(struct tcf_proto *tp, unsigned long fh,
struct sk_buff *skb, struct tcmsg *t)
{
struct tc_u_knode *n = (struct tc_u_knode*)fh;
unsigned char *b = skb->tail;
struct rtattr *rta;
if (n == NULL)
return skb->len;
t->tcm_handle = n->handle;
rta = (struct rtattr*)b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
if (TC_U32_KEY(n->handle) == 0) {
struct tc_u_hnode *ht = (struct tc_u_hnode*)fh;
u32 divisor = ht->divisor+1;
RTA_PUT(skb, TCA_U32_DIVISOR, 4, &divisor);
} else {
RTA_PUT(skb, TCA_U32_SEL,
sizeof(n->sel) + n->sel.nkeys*sizeof(struct tc_u32_key),
&n->sel);
if (n->ht_up) {
u32 htid = n->handle & 0xFFFFF000;
RTA_PUT(skb, TCA_U32_HASH, 4, &htid);
}
if (n->res.classid)
RTA_PUT(skb, TCA_U32_CLASSID, 4, &n->res.classid);
if (n->ht_down)
RTA_PUT(skb, TCA_U32_LINK, 4, &n->ht_down->handle);
#ifdef CONFIG_CLS_U32_MARK
if (n->mark.val || n->mark.mask)
RTA_PUT(skb, TCA_U32_MARK, sizeof(n->mark), &n->mark);
#endif
if (tcf_exts_dump(skb, &n->exts, &u32_ext_map) < 0)
goto rtattr_failure;
#ifdef CONFIG_NET_CLS_IND
if(strlen(n->indev))
RTA_PUT(skb, TCA_U32_INDEV, IFNAMSIZ, n->indev);
#endif
#ifdef CONFIG_CLS_U32_PERF
RTA_PUT(skb, TCA_U32_PCNT,
sizeof(struct tc_u32_pcnt) + n->sel.nkeys*sizeof(u64),
n->pf);
#endif
}
rta->rta_len = skb->tail - b;
if (TC_U32_KEY(n->handle))
if (tcf_exts_dump_stats(skb, &n->exts, &u32_ext_map) < 0)
goto rtattr_failure;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tcf_proto_ops cls_u32_ops = {
.next = NULL,
.kind = "u32",
.classify = u32_classify,
.init = u32_init,
.destroy = u32_destroy,
.get = u32_get,
.put = u32_put,
.change = u32_change,
.delete = u32_delete,
.walk = u32_walk,
.dump = u32_dump,
.owner = THIS_MODULE,
};
static int __init init_u32(void)
{
printk("u32 classifier\n");
#ifdef CONFIG_CLS_U32_PERF
printk(" Perfomance counters on\n");
#endif
#ifdef CONFIG_NET_CLS_POLICE
printk(" OLD policer on \n");
#endif
#ifdef CONFIG_NET_CLS_IND
printk(" input device check on \n");
#endif
#ifdef CONFIG_NET_CLS_ACT
printk(" Actions configured \n");
#endif
return register_tcf_proto_ops(&cls_u32_ops);
}
static void __exit exit_u32(void)
{
unregister_tcf_proto_ops(&cls_u32_ops);
}
module_init(init_u32)
module_exit(exit_u32)
MODULE_LICENSE("GPL");

101
net/sched/em_cmp.c Normal file
View File

@@ -0,0 +1,101 @@
/*
* net/sched/em_cmp.c Simple packet data comparison ematch
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Thomas Graf <tgraf@suug.ch>
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/skbuff.h>
#include <linux/tc_ematch/tc_em_cmp.h>
#include <net/pkt_cls.h>
static inline int cmp_needs_transformation(struct tcf_em_cmp *cmp)
{
return unlikely(cmp->flags & TCF_EM_CMP_TRANS);
}
static int em_cmp_match(struct sk_buff *skb, struct tcf_ematch *em,
struct tcf_pkt_info *info)
{
struct tcf_em_cmp *cmp = (struct tcf_em_cmp *) em->data;
unsigned char *ptr = tcf_get_base_ptr(skb, cmp->layer) + cmp->off;
u32 val = 0;
if (!tcf_valid_offset(skb, ptr, cmp->align))
return 0;
switch (cmp->align) {
case TCF_EM_ALIGN_U8:
val = *ptr;
break;
case TCF_EM_ALIGN_U16:
val = *ptr << 8;
val |= *(ptr+1);
if (cmp_needs_transformation(cmp))
val = be16_to_cpu(val);
break;
case TCF_EM_ALIGN_U32:
/* Worth checking boundries? The branching seems
* to get worse. Visit again. */
val = *ptr << 24;
val |= *(ptr+1) << 16;
val |= *(ptr+2) << 8;
val |= *(ptr+3);
if (cmp_needs_transformation(cmp))
val = be32_to_cpu(val);
break;
default:
return 0;
}
if (cmp->mask)
val &= cmp->mask;
switch (cmp->opnd) {
case TCF_EM_OPND_EQ:
return val == cmp->val;
case TCF_EM_OPND_LT:
return val < cmp->val;
case TCF_EM_OPND_GT:
return val > cmp->val;
}
return 0;
}
static struct tcf_ematch_ops em_cmp_ops = {
.kind = TCF_EM_CMP,
.datalen = sizeof(struct tcf_em_cmp),
.match = em_cmp_match,
.owner = THIS_MODULE,
.link = LIST_HEAD_INIT(em_cmp_ops.link)
};
static int __init init_em_cmp(void)
{
return tcf_em_register(&em_cmp_ops);
}
static void __exit exit_em_cmp(void)
{
tcf_em_unregister(&em_cmp_ops);
}
MODULE_LICENSE("GPL");
module_init(init_em_cmp);
module_exit(exit_em_cmp);

661
net/sched/em_meta.c Normal file
View File

@@ -0,0 +1,661 @@
/*
* net/sched/em_meta.c Metadata ematch
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Thomas Graf <tgraf@suug.ch>
*
* ==========================================================================
*
* The metadata ematch compares two meta objects where each object
* represents either a meta value stored in the kernel or a static
* value provided by userspace. The objects are not provided by
* userspace itself but rather a definition providing the information
* to build them. Every object is of a certain type which must be
* equal to the object it is being compared to.
*
* The definition of a objects conists of the type (meta type), a
* identifier (meta id) and additional type specific information.
* The meta id is either TCF_META_TYPE_VALUE for values provided by
* userspace or a index to the meta operations table consisting of
* function pointers to type specific meta data collectors returning
* the value of the requested meta value.
*
* lvalue rvalue
* +-----------+ +-----------+
* | type: INT | | type: INT |
* def | id: INDEV | | id: VALUE |
* | data: | | data: 3 |
* +-----------+ +-----------+
* | |
* ---> meta_ops[INT][INDEV](...) |
* | |
* ----------- |
* V V
* +-----------+ +-----------+
* | type: INT | | type: INT |
* obj | id: INDEV | | id: VALUE |
* | data: 2 |<--data got filled out | data: 3 |
* +-----------+ +-----------+
* | |
* --------------> 2 equals 3 <--------------
*
* This is a simplified schema, the complexity varies depending
* on the meta type. Obviously, the length of the data must also
* be provided for non-numeric types.
*
* Additionaly, type dependant modifiers such as shift operators
* or mask may be applied to extend the functionaliy. As of now,
* the variable length type supports shifting the byte string to
* the right, eating up any number of octets and thus supporting
* wildcard interface name comparisons such as "ppp%" matching
* ppp0..9.
*
* NOTE: Certain meta values depend on other subsystems and are
* only available if that subsytem is enabled in the kernel.
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/skbuff.h>
#include <linux/random.h>
#include <linux/tc_ematch/tc_em_meta.h>
#include <net/dst.h>
#include <net/route.h>
#include <net/pkt_cls.h>
struct meta_obj
{
unsigned long value;
unsigned int len;
};
struct meta_value
{
struct tcf_meta_val hdr;
unsigned long val;
unsigned int len;
};
struct meta_match
{
struct meta_value lvalue;
struct meta_value rvalue;
};
static inline int meta_id(struct meta_value *v)
{
return TCF_META_ID(v->hdr.kind);
}
static inline int meta_type(struct meta_value *v)
{
return TCF_META_TYPE(v->hdr.kind);
}
#define META_COLLECTOR(FUNC) static void meta_##FUNC(struct sk_buff *skb, \
struct tcf_pkt_info *info, struct meta_value *v, \
struct meta_obj *dst, int *err)
/**************************************************************************
* System status & misc
**************************************************************************/
META_COLLECTOR(int_random)
{
get_random_bytes(&dst->value, sizeof(dst->value));
}
static inline unsigned long fixed_loadavg(int load)
{
int rnd_load = load + (FIXED_1/200);
int rnd_frac = ((rnd_load & (FIXED_1-1)) * 100) >> FSHIFT;
return ((rnd_load >> FSHIFT) * 100) + rnd_frac;
}
META_COLLECTOR(int_loadavg_0)
{
dst->value = fixed_loadavg(avenrun[0]);
}
META_COLLECTOR(int_loadavg_1)
{
dst->value = fixed_loadavg(avenrun[1]);
}
META_COLLECTOR(int_loadavg_2)
{
dst->value = fixed_loadavg(avenrun[2]);
}
/**************************************************************************
* Device names & indices
**************************************************************************/
static inline int int_dev(struct net_device *dev, struct meta_obj *dst)
{
if (unlikely(dev == NULL))
return -1;
dst->value = dev->ifindex;
return 0;
}
static inline int var_dev(struct net_device *dev, struct meta_obj *dst)
{
if (unlikely(dev == NULL))
return -1;
dst->value = (unsigned long) dev->name;
dst->len = strlen(dev->name);
return 0;
}
META_COLLECTOR(int_dev)
{
*err = int_dev(skb->dev, dst);
}
META_COLLECTOR(var_dev)
{
*err = var_dev(skb->dev, dst);
}
META_COLLECTOR(int_indev)
{
*err = int_dev(skb->input_dev, dst);
}
META_COLLECTOR(var_indev)
{
*err = var_dev(skb->input_dev, dst);
}
META_COLLECTOR(int_realdev)
{
*err = int_dev(skb->real_dev, dst);
}
META_COLLECTOR(var_realdev)
{
*err = var_dev(skb->real_dev, dst);
}
/**************************************************************************
* skb attributes
**************************************************************************/
META_COLLECTOR(int_priority)
{
dst->value = skb->priority;
}
META_COLLECTOR(int_protocol)
{
/* Let userspace take care of the byte ordering */
dst->value = skb->protocol;
}
META_COLLECTOR(int_security)
{
dst->value = skb->security;
}
META_COLLECTOR(int_pkttype)
{
dst->value = skb->pkt_type;
}
META_COLLECTOR(int_pktlen)
{
dst->value = skb->len;
}
META_COLLECTOR(int_datalen)
{
dst->value = skb->data_len;
}
META_COLLECTOR(int_maclen)
{
dst->value = skb->mac_len;
}
/**************************************************************************
* Netfilter
**************************************************************************/
#ifdef CONFIG_NETFILTER
META_COLLECTOR(int_nfmark)
{
dst->value = skb->nfmark;
}
#endif
/**************************************************************************
* Traffic Control
**************************************************************************/
META_COLLECTOR(int_tcindex)
{
dst->value = skb->tc_index;
}
#ifdef CONFIG_NET_CLS_ACT
META_COLLECTOR(int_tcverd)
{
dst->value = skb->tc_verd;
}
META_COLLECTOR(int_tcclassid)
{
dst->value = skb->tc_classid;
}
#endif
/**************************************************************************
* Routing
**************************************************************************/
#ifdef CONFIG_NET_CLS_ROUTE
META_COLLECTOR(int_rtclassid)
{
if (unlikely(skb->dst == NULL))
*err = -1;
else
dst->value = skb->dst->tclassid;
}
#endif
META_COLLECTOR(int_rtiif)
{
if (unlikely(skb->dst == NULL))
*err = -1;
else
dst->value = ((struct rtable*) skb->dst)->fl.iif;
}
/**************************************************************************
* Meta value collectors assignment table
**************************************************************************/
struct meta_ops
{
void (*get)(struct sk_buff *, struct tcf_pkt_info *,
struct meta_value *, struct meta_obj *, int *);
};
/* Meta value operations table listing all meta value collectors and
* assigns them to a type and meta id. */
static struct meta_ops __meta_ops[TCF_META_TYPE_MAX+1][TCF_META_ID_MAX+1] = {
[TCF_META_TYPE_VAR] = {
[TCF_META_ID_DEV] = { .get = meta_var_dev },
[TCF_META_ID_INDEV] = { .get = meta_var_indev },
[TCF_META_ID_REALDEV] = { .get = meta_var_realdev }
},
[TCF_META_TYPE_INT] = {
[TCF_META_ID_RANDOM] = { .get = meta_int_random },
[TCF_META_ID_LOADAVG_0] = { .get = meta_int_loadavg_0 },
[TCF_META_ID_LOADAVG_1] = { .get = meta_int_loadavg_1 },
[TCF_META_ID_LOADAVG_2] = { .get = meta_int_loadavg_2 },
[TCF_META_ID_DEV] = { .get = meta_int_dev },
[TCF_META_ID_INDEV] = { .get = meta_int_indev },
[TCF_META_ID_REALDEV] = { .get = meta_int_realdev },
[TCF_META_ID_PRIORITY] = { .get = meta_int_priority },
[TCF_META_ID_PROTOCOL] = { .get = meta_int_protocol },
[TCF_META_ID_SECURITY] = { .get = meta_int_security },
[TCF_META_ID_PKTTYPE] = { .get = meta_int_pkttype },
[TCF_META_ID_PKTLEN] = { .get = meta_int_pktlen },
[TCF_META_ID_DATALEN] = { .get = meta_int_datalen },
[TCF_META_ID_MACLEN] = { .get = meta_int_maclen },
#ifdef CONFIG_NETFILTER
[TCF_META_ID_NFMARK] = { .get = meta_int_nfmark },
#endif
[TCF_META_ID_TCINDEX] = { .get = meta_int_tcindex },
#ifdef CONFIG_NET_CLS_ACT
[TCF_META_ID_TCVERDICT] = { .get = meta_int_tcverd },
[TCF_META_ID_TCCLASSID] = { .get = meta_int_tcclassid },
#endif
#ifdef CONFIG_NET_CLS_ROUTE
[TCF_META_ID_RTCLASSID] = { .get = meta_int_rtclassid },
#endif
[TCF_META_ID_RTIIF] = { .get = meta_int_rtiif }
}
};
static inline struct meta_ops * meta_ops(struct meta_value *val)
{
return &__meta_ops[meta_type(val)][meta_id(val)];
}
/**************************************************************************
* Type specific operations for TCF_META_TYPE_VAR
**************************************************************************/
static int meta_var_compare(struct meta_obj *a, struct meta_obj *b)
{
int r = a->len - b->len;
if (r == 0)
r = memcmp((void *) a->value, (void *) b->value, a->len);
return r;
}
static int meta_var_change(struct meta_value *dst, struct rtattr *rta)
{
int len = RTA_PAYLOAD(rta);
dst->val = (unsigned long) kmalloc(len, GFP_KERNEL);
if (dst->val == 0UL)
return -ENOMEM;
memcpy((void *) dst->val, RTA_DATA(rta), len);
dst->len = len;
return 0;
}
static void meta_var_destroy(struct meta_value *v)
{
if (v->val)
kfree((void *) v->val);
}
static void meta_var_apply_extras(struct meta_value *v,
struct meta_obj *dst)
{
int shift = v->hdr.shift;
if (shift && shift < dst->len)
dst->len -= shift;
}
static int meta_var_dump(struct sk_buff *skb, struct meta_value *v, int tlv)
{
if (v->val && v->len)
RTA_PUT(skb, tlv, v->len, (void *) v->val);
return 0;
rtattr_failure:
return -1;
}
/**************************************************************************
* Type specific operations for TCF_META_TYPE_INT
**************************************************************************/
static int meta_int_compare(struct meta_obj *a, struct meta_obj *b)
{
/* Let gcc optimize it, the unlikely is not really based on
* some numbers but jump free code for mismatches seems
* more logical. */
if (unlikely(a == b))
return 0;
else if (a < b)
return -1;
else
return 1;
}
static int meta_int_change(struct meta_value *dst, struct rtattr *rta)
{
if (RTA_PAYLOAD(rta) >= sizeof(unsigned long)) {
dst->val = *(unsigned long *) RTA_DATA(rta);
dst->len = sizeof(unsigned long);
} else if (RTA_PAYLOAD(rta) == sizeof(u32)) {
dst->val = *(u32 *) RTA_DATA(rta);
dst->len = sizeof(u32);
} else
return -EINVAL;
return 0;
}
static void meta_int_apply_extras(struct meta_value *v,
struct meta_obj *dst)
{
if (v->hdr.shift)
dst->value >>= v->hdr.shift;
if (v->val)
dst->value &= v->val;
}
static int meta_int_dump(struct sk_buff *skb, struct meta_value *v, int tlv)
{
if (v->len == sizeof(unsigned long))
RTA_PUT(skb, tlv, sizeof(unsigned long), &v->val);
else if (v->len == sizeof(u32)) {
u32 d = v->val;
RTA_PUT(skb, tlv, sizeof(d), &d);
}
return 0;
rtattr_failure:
return -1;
}
/**************************************************************************
* Type specific operations table
**************************************************************************/
struct meta_type_ops
{
void (*destroy)(struct meta_value *);
int (*compare)(struct meta_obj *, struct meta_obj *);
int (*change)(struct meta_value *, struct rtattr *);
void (*apply_extras)(struct meta_value *, struct meta_obj *);
int (*dump)(struct sk_buff *, struct meta_value *, int);
};
static struct meta_type_ops __meta_type_ops[TCF_META_TYPE_MAX+1] = {
[TCF_META_TYPE_VAR] = {
.destroy = meta_var_destroy,
.compare = meta_var_compare,
.change = meta_var_change,
.apply_extras = meta_var_apply_extras,
.dump = meta_var_dump
},
[TCF_META_TYPE_INT] = {
.compare = meta_int_compare,
.change = meta_int_change,
.apply_extras = meta_int_apply_extras,
.dump = meta_int_dump
}
};
static inline struct meta_type_ops * meta_type_ops(struct meta_value *v)
{
return &__meta_type_ops[meta_type(v)];
}
/**************************************************************************
* Core
**************************************************************************/
static inline int meta_get(struct sk_buff *skb, struct tcf_pkt_info *info,
struct meta_value *v, struct meta_obj *dst)
{
int err = 0;
if (meta_id(v) == TCF_META_ID_VALUE) {
dst->value = v->val;
dst->len = v->len;
return 0;
}
meta_ops(v)->get(skb, info, v, dst, &err);
if (err < 0)
return err;
if (meta_type_ops(v)->apply_extras)
meta_type_ops(v)->apply_extras(v, dst);
return 0;
}
static int em_meta_match(struct sk_buff *skb, struct tcf_ematch *m,
struct tcf_pkt_info *info)
{
int r;
struct meta_match *meta = (struct meta_match *) m->data;
struct meta_obj l_value, r_value;
if (meta_get(skb, info, &meta->lvalue, &l_value) < 0 ||
meta_get(skb, info, &meta->rvalue, &r_value) < 0)
return 0;
r = meta_type_ops(&meta->lvalue)->compare(&l_value, &r_value);
switch (meta->lvalue.hdr.op) {
case TCF_EM_OPND_EQ:
return !r;
case TCF_EM_OPND_LT:
return r < 0;
case TCF_EM_OPND_GT:
return r > 0;
}
return 0;
}
static inline void meta_delete(struct meta_match *meta)
{
struct meta_type_ops *ops = meta_type_ops(&meta->lvalue);
if (ops && ops->destroy) {
ops->destroy(&meta->lvalue);
ops->destroy(&meta->rvalue);
}
kfree(meta);
}
static inline int meta_change_data(struct meta_value *dst, struct rtattr *rta)
{
if (rta) {
if (RTA_PAYLOAD(rta) == 0)
return -EINVAL;
return meta_type_ops(dst)->change(dst, rta);
}
return 0;
}
static inline int meta_is_supported(struct meta_value *val)
{
return (!meta_id(val) || meta_ops(val)->get);
}
static int em_meta_change(struct tcf_proto *tp, void *data, int len,
struct tcf_ematch *m)
{
int err = -EINVAL;
struct rtattr *tb[TCA_EM_META_MAX];
struct tcf_meta_hdr *hdr;
struct meta_match *meta = NULL;
if (rtattr_parse(tb, TCA_EM_META_MAX, data, len) < 0)
goto errout;
if (tb[TCA_EM_META_HDR-1] == NULL ||
RTA_PAYLOAD(tb[TCA_EM_META_HDR-1]) < sizeof(*hdr))
goto errout;
hdr = RTA_DATA(tb[TCA_EM_META_HDR-1]);
if (TCF_META_TYPE(hdr->left.kind) != TCF_META_TYPE(hdr->right.kind) ||
TCF_META_TYPE(hdr->left.kind) > TCF_META_TYPE_MAX ||
TCF_META_ID(hdr->left.kind) > TCF_META_ID_MAX ||
TCF_META_ID(hdr->right.kind) > TCF_META_ID_MAX)
goto errout;
meta = kmalloc(sizeof(*meta), GFP_KERNEL);
if (meta == NULL)
goto errout;
memset(meta, 0, sizeof(*meta));
memcpy(&meta->lvalue.hdr, &hdr->left, sizeof(hdr->left));
memcpy(&meta->rvalue.hdr, &hdr->right, sizeof(hdr->right));
if (!meta_is_supported(&meta->lvalue) ||
!meta_is_supported(&meta->rvalue)) {
err = -EOPNOTSUPP;
goto errout;
}
if (meta_change_data(&meta->lvalue, tb[TCA_EM_META_LVALUE-1]) < 0 ||
meta_change_data(&meta->rvalue, tb[TCA_EM_META_RVALUE-1]) < 0)
goto errout;
m->datalen = sizeof(*meta);
m->data = (unsigned long) meta;
err = 0;
errout:
if (err && meta)
meta_delete(meta);
return err;
}
static void em_meta_destroy(struct tcf_proto *tp, struct tcf_ematch *m)
{
if (m)
meta_delete((struct meta_match *) m->data);
}
static int em_meta_dump(struct sk_buff *skb, struct tcf_ematch *em)
{
struct meta_match *meta = (struct meta_match *) em->data;
struct tcf_meta_hdr hdr;
struct meta_type_ops *ops;
memset(&hdr, 0, sizeof(hdr));
memcpy(&hdr.left, &meta->lvalue.hdr, sizeof(hdr.left));
memcpy(&hdr.right, &meta->rvalue.hdr, sizeof(hdr.right));
RTA_PUT(skb, TCA_EM_META_HDR, sizeof(hdr), &hdr);
ops = meta_type_ops(&meta->lvalue);
if (ops->dump(skb, &meta->lvalue, TCA_EM_META_LVALUE) < 0 ||
ops->dump(skb, &meta->rvalue, TCA_EM_META_RVALUE) < 0)
goto rtattr_failure;
return 0;
rtattr_failure:
return -1;
}
static struct tcf_ematch_ops em_meta_ops = {
.kind = TCF_EM_META,
.change = em_meta_change,
.match = em_meta_match,
.destroy = em_meta_destroy,
.dump = em_meta_dump,
.owner = THIS_MODULE,
.link = LIST_HEAD_INIT(em_meta_ops.link)
};
static int __init init_em_meta(void)
{
return tcf_em_register(&em_meta_ops);
}
static void __exit exit_em_meta(void)
{
tcf_em_unregister(&em_meta_ops);
}
MODULE_LICENSE("GPL");
module_init(init_em_meta);
module_exit(exit_em_meta);

82
net/sched/em_nbyte.c Normal file
View File

@@ -0,0 +1,82 @@
/*
* net/sched/em_nbyte.c N-Byte ematch
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Thomas Graf <tgraf@suug.ch>
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/skbuff.h>
#include <linux/tc_ematch/tc_em_nbyte.h>
#include <net/pkt_cls.h>
struct nbyte_data
{
struct tcf_em_nbyte hdr;
char pattern[0];
};
static int em_nbyte_change(struct tcf_proto *tp, void *data, int data_len,
struct tcf_ematch *em)
{
struct tcf_em_nbyte *nbyte = data;
if (data_len < sizeof(*nbyte) ||
data_len < (sizeof(*nbyte) + nbyte->len))
return -EINVAL;
em->datalen = sizeof(*nbyte) + nbyte->len;
em->data = (unsigned long) kmalloc(em->datalen, GFP_KERNEL);
if (em->data == 0UL)
return -ENOBUFS;
memcpy((void *) em->data, data, em->datalen);
return 0;
}
static int em_nbyte_match(struct sk_buff *skb, struct tcf_ematch *em,
struct tcf_pkt_info *info)
{
struct nbyte_data *nbyte = (struct nbyte_data *) em->data;
unsigned char *ptr = tcf_get_base_ptr(skb, nbyte->hdr.layer);
ptr += nbyte->hdr.off;
if (!tcf_valid_offset(skb, ptr, nbyte->hdr.len))
return 0;
return !memcmp(ptr + nbyte->hdr.off, nbyte->pattern, nbyte->hdr.len);
}
static struct tcf_ematch_ops em_nbyte_ops = {
.kind = TCF_EM_NBYTE,
.change = em_nbyte_change,
.match = em_nbyte_match,
.owner = THIS_MODULE,
.link = LIST_HEAD_INIT(em_nbyte_ops.link)
};
static int __init init_em_nbyte(void)
{
return tcf_em_register(&em_nbyte_ops);
}
static void __exit exit_em_nbyte(void)
{
tcf_em_unregister(&em_nbyte_ops);
}
MODULE_LICENSE("GPL");
module_init(init_em_nbyte);
module_exit(exit_em_nbyte);

63
net/sched/em_u32.c Normal file
View File

@@ -0,0 +1,63 @@
/*
* net/sched/em_u32.c U32 Ematch
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Thomas Graf <tgraf@suug.ch>
* Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
* Based on net/sched/cls_u32.c
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/skbuff.h>
#include <net/pkt_cls.h>
static int em_u32_match(struct sk_buff *skb, struct tcf_ematch *em,
struct tcf_pkt_info *info)
{
struct tc_u32_key *key = (struct tc_u32_key *) em->data;
unsigned char *ptr = skb->nh.raw;
if (info) {
if (info->ptr)
ptr = info->ptr;
ptr += (info->nexthdr & key->offmask);
}
ptr += key->off;
if (!tcf_valid_offset(skb, ptr, sizeof(u32)))
return 0;
return !(((*(u32*) ptr) ^ key->val) & key->mask);
}
static struct tcf_ematch_ops em_u32_ops = {
.kind = TCF_EM_U32,
.datalen = sizeof(struct tc_u32_key),
.match = em_u32_match,
.owner = THIS_MODULE,
.link = LIST_HEAD_INIT(em_u32_ops.link)
};
static int __init init_em_u32(void)
{
return tcf_em_register(&em_u32_ops);
}
static void __exit exit_em_u32(void)
{
tcf_em_unregister(&em_u32_ops);
}
MODULE_LICENSE("GPL");
module_init(init_em_u32);
module_exit(exit_em_u32);

524
net/sched/ematch.c Normal file
View File

@@ -0,0 +1,524 @@
/*
* net/sched/ematch.c Extended Match API
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Thomas Graf <tgraf@suug.ch>
*
* ==========================================================================
*
* An extended match (ematch) is a small classification tool not worth
* writing a full classifier for. Ematches can be interconnected to form
* a logic expression and get attached to classifiers to extend their
* functionatlity.
*
* The userspace part transforms the logic expressions into an array
* consisting of multiple sequences of interconnected ematches separated
* by markers. Precedence is implemented by a special ematch kind
* referencing a sequence beyond the marker of the current sequence
* causing the current position in the sequence to be pushed onto a stack
* to allow the current position to be overwritten by the position referenced
* in the special ematch. Matching continues in the new sequence until a
* marker is reached causing the position to be restored from the stack.
*
* Example:
* A AND (B1 OR B2) AND C AND D
*
* ------->-PUSH-------
* -->-- / -->-- \ -->--
* / \ / / \ \ / \
* +-------+-------+-------+-------+-------+--------+
* | A AND | B AND | C AND | D END | B1 OR | B2 END |
* +-------+-------+-------+-------+-------+--------+
* \ /
* --------<-POP---------
*
* where B is a virtual ematch referencing to sequence starting with B1.
*
* ==========================================================================
*
* How to write an ematch in 60 seconds
* ------------------------------------
*
* 1) Provide a matcher function:
* static int my_match(struct sk_buff *skb, struct tcf_ematch *m,
* struct tcf_pkt_info *info)
* {
* struct mydata *d = (struct mydata *) m->data;
*
* if (...matching goes here...)
* return 1;
* else
* return 0;
* }
*
* 2) Fill out a struct tcf_ematch_ops:
* static struct tcf_ematch_ops my_ops = {
* .kind = unique id,
* .datalen = sizeof(struct mydata),
* .match = my_match,
* .owner = THIS_MODULE,
* };
*
* 3) Register/Unregister your ematch:
* static int __init init_my_ematch(void)
* {
* return tcf_em_register(&my_ops);
* }
*
* static void __exit exit_my_ematch(void)
* {
* return tcf_em_unregister(&my_ops);
* }
*
* module_init(init_my_ematch);
* module_exit(exit_my_ematch);
*
* 4) By now you should have two more seconds left, barely enough to
* open up a beer to watch the compilation going.
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/rtnetlink.h>
#include <linux/skbuff.h>
#include <net/pkt_cls.h>
#include <config/net/ematch/stack.h>
static LIST_HEAD(ematch_ops);
static DEFINE_RWLOCK(ematch_mod_lock);
static inline struct tcf_ematch_ops * tcf_em_lookup(u16 kind)
{
struct tcf_ematch_ops *e = NULL;
read_lock(&ematch_mod_lock);
list_for_each_entry(e, &ematch_ops, link) {
if (kind == e->kind) {
if (!try_module_get(e->owner))
e = NULL;
read_unlock(&ematch_mod_lock);
return e;
}
}
read_unlock(&ematch_mod_lock);
return NULL;
}
/**
* tcf_em_register - register an extended match
*
* @ops: ematch operations lookup table
*
* This function must be called by ematches to announce their presence.
* The given @ops must have kind set to a unique identifier and the
* callback match() must be implemented. All other callbacks are optional
* and a fallback implementation is used instead.
*
* Returns -EEXISTS if an ematch of the same kind has already registered.
*/
int tcf_em_register(struct tcf_ematch_ops *ops)
{
int err = -EEXIST;
struct tcf_ematch_ops *e;
if (ops->match == NULL)
return -EINVAL;
write_lock(&ematch_mod_lock);
list_for_each_entry(e, &ematch_ops, link)
if (ops->kind == e->kind)
goto errout;
list_add_tail(&ops->link, &ematch_ops);
err = 0;
errout:
write_unlock(&ematch_mod_lock);
return err;
}
/**
* tcf_em_unregister - unregster and extended match
*
* @ops: ematch operations lookup table
*
* This function must be called by ematches to announce their disappearance
* for examples when the module gets unloaded. The @ops parameter must be
* the same as the one used for registration.
*
* Returns -ENOENT if no matching ematch was found.
*/
int tcf_em_unregister(struct tcf_ematch_ops *ops)
{
int err = 0;
struct tcf_ematch_ops *e;
write_lock(&ematch_mod_lock);
list_for_each_entry(e, &ematch_ops, link) {
if (e == ops) {
list_del(&e->link);
goto out;
}
}
err = -ENOENT;
out:
write_unlock(&ematch_mod_lock);
return err;
}
static inline struct tcf_ematch * tcf_em_get_match(struct tcf_ematch_tree *tree,
int index)
{
return &tree->matches[index];
}
static int tcf_em_validate(struct tcf_proto *tp,
struct tcf_ematch_tree_hdr *tree_hdr,
struct tcf_ematch *em, struct rtattr *rta, int idx)
{
int err = -EINVAL;
struct tcf_ematch_hdr *em_hdr = RTA_DATA(rta);
int data_len = RTA_PAYLOAD(rta) - sizeof(*em_hdr);
void *data = (void *) em_hdr + sizeof(*em_hdr);
if (!TCF_EM_REL_VALID(em_hdr->flags))
goto errout;
if (em_hdr->kind == TCF_EM_CONTAINER) {
/* Special ematch called "container", carries an index
* referencing an external ematch sequence. */
u32 ref;
if (data_len < sizeof(ref))
goto errout;
ref = *(u32 *) data;
if (ref >= tree_hdr->nmatches)
goto errout;
/* We do not allow backward jumps to avoid loops and jumps
* to our own position are of course illegal. */
if (ref <= idx)
goto errout;
em->data = ref;
} else {
/* Note: This lookup will increase the module refcnt
* of the ematch module referenced. In case of a failure,
* a destroy function is called by the underlying layer
* which automatically releases the reference again, therefore
* the module MUST not be given back under any circumstances
* here. Be aware, the destroy function assumes that the
* module is held if the ops field is non zero. */
em->ops = tcf_em_lookup(em_hdr->kind);
if (em->ops == NULL) {
err = -ENOENT;
goto errout;
}
/* ematch module provides expected length of data, so we
* can do a basic sanity check. */
if (em->ops->datalen && data_len < em->ops->datalen)
goto errout;
if (em->ops->change) {
err = em->ops->change(tp, data, data_len, em);
if (err < 0)
goto errout;
} else if (data_len > 0) {
/* ematch module doesn't provide an own change
* procedure and expects us to allocate and copy
* the ematch data.
*
* TCF_EM_SIMPLE may be specified stating that the
* data only consists of a u32 integer and the module
* does not expected a memory reference but rather
* the value carried. */
if (em_hdr->flags & TCF_EM_SIMPLE) {
if (data_len < sizeof(u32))
goto errout;
em->data = *(u32 *) data;
} else {
void *v = kmalloc(data_len, GFP_KERNEL);
if (v == NULL) {
err = -ENOBUFS;
goto errout;
}
memcpy(v, data, data_len);
em->data = (unsigned long) v;
}
}
}
em->matchid = em_hdr->matchid;
em->flags = em_hdr->flags;
em->datalen = data_len;
err = 0;
errout:
return err;
}
/**
* tcf_em_tree_validate - validate ematch config TLV and build ematch tree
*
* @tp: classifier kind handle
* @rta: ematch tree configuration TLV
* @tree: destination ematch tree variable to store the resulting
* ematch tree.
*
* This function validates the given configuration TLV @rta and builds an
* ematch tree in @tree. The resulting tree must later be copied into
* the private classifier data using tcf_em_tree_change(). You MUST NOT
* provide the ematch tree variable of the private classifier data directly,
* the changes would not be locked properly.
*
* Returns a negative error code if the configuration TLV contains errors.
*/
int tcf_em_tree_validate(struct tcf_proto *tp, struct rtattr *rta,
struct tcf_ematch_tree *tree)
{
int idx, list_len, matches_len, err = -EINVAL;
struct rtattr *tb[TCA_EMATCH_TREE_MAX];
struct rtattr *rt_match, *rt_hdr, *rt_list;
struct tcf_ematch_tree_hdr *tree_hdr;
struct tcf_ematch *em;
if (rtattr_parse_nested(tb, TCA_EMATCH_TREE_MAX, rta) < 0)
goto errout;
rt_hdr = tb[TCA_EMATCH_TREE_HDR-1];
rt_list = tb[TCA_EMATCH_TREE_LIST-1];
if (rt_hdr == NULL || rt_list == NULL)
goto errout;
if (RTA_PAYLOAD(rt_hdr) < sizeof(*tree_hdr) ||
RTA_PAYLOAD(rt_list) < sizeof(*rt_match))
goto errout;
tree_hdr = RTA_DATA(rt_hdr);
memcpy(&tree->hdr, tree_hdr, sizeof(*tree_hdr));
rt_match = RTA_DATA(rt_list);
list_len = RTA_PAYLOAD(rt_list);
matches_len = tree_hdr->nmatches * sizeof(*em);
tree->matches = kmalloc(matches_len, GFP_KERNEL);
if (tree->matches == NULL)
goto errout;
memset(tree->matches, 0, matches_len);
/* We do not use rtattr_parse_nested here because the maximum
* number of attributes is unknown. This saves us the allocation
* for a tb buffer which would serve no purpose at all.
*
* The array of rt attributes is parsed in the order as they are
* provided, their type must be incremental from 1 to n. Even
* if it does not serve any real purpose, a failure of sticking
* to this policy will result in parsing failure. */
for (idx = 0; RTA_OK(rt_match, list_len); idx++) {
err = -EINVAL;
if (rt_match->rta_type != (idx + 1))
goto errout_abort;
if (idx >= tree_hdr->nmatches)
goto errout_abort;
if (RTA_PAYLOAD(rt_match) < sizeof(struct tcf_ematch_hdr))
goto errout_abort;
em = tcf_em_get_match(tree, idx);
err = tcf_em_validate(tp, tree_hdr, em, rt_match, idx);
if (err < 0)
goto errout_abort;
rt_match = RTA_NEXT(rt_match, list_len);
}
/* Check if the number of matches provided by userspace actually
* complies with the array of matches. The number was used for
* the validation of references and a mismatch could lead to
* undefined references during the matching process. */
if (idx != tree_hdr->nmatches) {
err = -EINVAL;
goto errout_abort;
}
err = 0;
errout:
return err;
errout_abort:
tcf_em_tree_destroy(tp, tree);
return err;
}
/**
* tcf_em_tree_destroy - destroy an ematch tree
*
* @tp: classifier kind handle
* @tree: ematch tree to be deleted
*
* This functions destroys an ematch tree previously created by
* tcf_em_tree_validate()/tcf_em_tree_change(). You must ensure that
* the ematch tree is not in use before calling this function.
*/
void tcf_em_tree_destroy(struct tcf_proto *tp, struct tcf_ematch_tree *tree)
{
int i;
if (tree->matches == NULL)
return;
for (i = 0; i < tree->hdr.nmatches; i++) {
struct tcf_ematch *em = tcf_em_get_match(tree, i);
if (em->ops) {
if (em->ops->destroy)
em->ops->destroy(tp, em);
else if (!tcf_em_is_simple(em) && em->data)
kfree((void *) em->data);
module_put(em->ops->owner);
}
}
tree->hdr.nmatches = 0;
kfree(tree->matches);
}
/**
* tcf_em_tree_dump - dump ematch tree into a rtnl message
*
* @skb: skb holding the rtnl message
* @t: ematch tree to be dumped
* @tlv: TLV type to be used to encapsulate the tree
*
* This function dumps a ematch tree into a rtnl message. It is valid to
* call this function while the ematch tree is in use.
*
* Returns -1 if the skb tailroom is insufficient.
*/
int tcf_em_tree_dump(struct sk_buff *skb, struct tcf_ematch_tree *tree, int tlv)
{
int i;
struct rtattr * top_start = (struct rtattr*) skb->tail;
struct rtattr * list_start;
RTA_PUT(skb, tlv, 0, NULL);
RTA_PUT(skb, TCA_EMATCH_TREE_HDR, sizeof(tree->hdr), &tree->hdr);
list_start = (struct rtattr *) skb->tail;
RTA_PUT(skb, TCA_EMATCH_TREE_LIST, 0, NULL);
for (i = 0; i < tree->hdr.nmatches; i++) {
struct rtattr *match_start = (struct rtattr*) skb->tail;
struct tcf_ematch *em = tcf_em_get_match(tree, i);
struct tcf_ematch_hdr em_hdr = {
.kind = em->ops ? em->ops->kind : TCF_EM_CONTAINER,
.matchid = em->matchid,
.flags = em->flags
};
RTA_PUT(skb, i+1, sizeof(em_hdr), &em_hdr);
if (em->ops && em->ops->dump) {
if (em->ops->dump(skb, em) < 0)
goto rtattr_failure;
} else if (tcf_em_is_container(em) || tcf_em_is_simple(em)) {
u32 u = em->data;
RTA_PUT_NOHDR(skb, sizeof(u), &u);
} else if (em->datalen > 0)
RTA_PUT_NOHDR(skb, em->datalen, (void *) em->data);
match_start->rta_len = skb->tail - (u8*) match_start;
}
list_start->rta_len = skb->tail - (u8 *) list_start;
top_start->rta_len = skb->tail - (u8 *) top_start;
return 0;
rtattr_failure:
return -1;
}
static inline int tcf_em_match(struct sk_buff *skb, struct tcf_ematch *em,
struct tcf_pkt_info *info)
{
int r = em->ops->match(skb, em, info);
return tcf_em_is_inverted(em) ? !r : r;
}
/* Do not use this function directly, use tcf_em_tree_match instead */
int __tcf_em_tree_match(struct sk_buff *skb, struct tcf_ematch_tree *tree,
struct tcf_pkt_info *info)
{
int stackp = 0, match_idx = 0, res = 0;
struct tcf_ematch *cur_match;
int stack[CONFIG_NET_EMATCH_STACK];
proceed:
while (match_idx < tree->hdr.nmatches) {
cur_match = tcf_em_get_match(tree, match_idx);
if (tcf_em_is_container(cur_match)) {
if (unlikely(stackp >= CONFIG_NET_EMATCH_STACK))
goto stack_overflow;
stack[stackp++] = match_idx;
match_idx = cur_match->data;
goto proceed;
}
res = tcf_em_match(skb, cur_match, info);
if (tcf_em_early_end(cur_match, res))
break;
match_idx++;
}
pop_stack:
if (stackp > 0) {
match_idx = stack[--stackp];
cur_match = tcf_em_get_match(tree, match_idx);
if (tcf_em_early_end(cur_match, res))
goto pop_stack;
else {
match_idx++;
goto proceed;
}
}
return res;
stack_overflow:
if (net_ratelimit())
printk("Local stack overflow, increase NET_EMATCH_STACK\n");
return -1;
}
EXPORT_SYMBOL(tcf_em_register);
EXPORT_SYMBOL(tcf_em_unregister);
EXPORT_SYMBOL(tcf_em_tree_validate);
EXPORT_SYMBOL(tcf_em_tree_destroy);
EXPORT_SYMBOL(tcf_em_tree_dump);
EXPORT_SYMBOL(__tcf_em_tree_match);

197
net/sched/estimator.c Normal file
View File

@@ -0,0 +1,197 @@
/*
* net/sched/estimator.c Simple rate estimator.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/jiffies.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/init.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
/*
This code is NOT intended to be used for statistics collection,
its purpose is to provide a base for statistical multiplexing
for controlled load service.
If you need only statistics, run a user level daemon which
periodically reads byte counters.
Unfortunately, rate estimation is not a very easy task.
F.e. I did not find a simple way to estimate the current peak rate
and even failed to formulate the problem 8)8)
So I preferred not to built an estimator into the scheduler,
but run this task separately.
Ideally, it should be kernel thread(s), but for now it runs
from timers, which puts apparent top bounds on the number of rated
flows, has minimal overhead on small, but is enough
to handle controlled load service, sets of aggregates.
We measure rate over A=(1<<interval) seconds and evaluate EWMA:
avrate = avrate*(1-W) + rate*W
where W is chosen as negative power of 2: W = 2^(-ewma_log)
The resulting time constant is:
T = A/(-ln(1-W))
NOTES.
* The stored value for avbps is scaled by 2^5, so that maximal
rate is ~1Gbit, avpps is scaled by 2^10.
* Minimal interval is HZ/4=250msec (it is the greatest common divisor
for HZ=100 and HZ=1024 8)), maximal interval
is (HZ*2^EST_MAX_INTERVAL)/4 = 8sec. Shorter intervals
are too expensive, longer ones can be implemented
at user level painlessly.
*/
#define EST_MAX_INTERVAL 5
struct qdisc_estimator
{
struct qdisc_estimator *next;
struct tc_stats *stats;
spinlock_t *stats_lock;
unsigned interval;
int ewma_log;
u64 last_bytes;
u32 last_packets;
u32 avpps;
u32 avbps;
};
struct qdisc_estimator_head
{
struct timer_list timer;
struct qdisc_estimator *list;
};
static struct qdisc_estimator_head elist[EST_MAX_INTERVAL+1];
/* Estimator array lock */
static DEFINE_RWLOCK(est_lock);
static void est_timer(unsigned long arg)
{
int idx = (int)arg;
struct qdisc_estimator *e;
read_lock(&est_lock);
for (e = elist[idx].list; e; e = e->next) {
struct tc_stats *st = e->stats;
u64 nbytes;
u32 npackets;
u32 rate;
spin_lock(e->stats_lock);
nbytes = st->bytes;
npackets = st->packets;
rate = (nbytes - e->last_bytes)<<(7 - idx);
e->last_bytes = nbytes;
e->avbps += ((long)rate - (long)e->avbps) >> e->ewma_log;
st->bps = (e->avbps+0xF)>>5;
rate = (npackets - e->last_packets)<<(12 - idx);
e->last_packets = npackets;
e->avpps += ((long)rate - (long)e->avpps) >> e->ewma_log;
e->stats->pps = (e->avpps+0x1FF)>>10;
spin_unlock(e->stats_lock);
}
mod_timer(&elist[idx].timer, jiffies + ((HZ<<idx)/4));
read_unlock(&est_lock);
}
int qdisc_new_estimator(struct tc_stats *stats, spinlock_t *stats_lock, struct rtattr *opt)
{
struct qdisc_estimator *est;
struct tc_estimator *parm = RTA_DATA(opt);
if (RTA_PAYLOAD(opt) < sizeof(*parm))
return -EINVAL;
if (parm->interval < -2 || parm->interval > 3)
return -EINVAL;
est = kmalloc(sizeof(*est), GFP_KERNEL);
if (est == NULL)
return -ENOBUFS;
memset(est, 0, sizeof(*est));
est->interval = parm->interval + 2;
est->stats = stats;
est->stats_lock = stats_lock;
est->ewma_log = parm->ewma_log;
est->last_bytes = stats->bytes;
est->avbps = stats->bps<<5;
est->last_packets = stats->packets;
est->avpps = stats->pps<<10;
est->next = elist[est->interval].list;
if (est->next == NULL) {
init_timer(&elist[est->interval].timer);
elist[est->interval].timer.data = est->interval;
elist[est->interval].timer.expires = jiffies + ((HZ<<est->interval)/4);
elist[est->interval].timer.function = est_timer;
add_timer(&elist[est->interval].timer);
}
write_lock_bh(&est_lock);
elist[est->interval].list = est;
write_unlock_bh(&est_lock);
return 0;
}
void qdisc_kill_estimator(struct tc_stats *stats)
{
int idx;
struct qdisc_estimator *est, **pest;
for (idx=0; idx <= EST_MAX_INTERVAL; idx++) {
int killed = 0;
pest = &elist[idx].list;
while ((est=*pest) != NULL) {
if (est->stats != stats) {
pest = &est->next;
continue;
}
write_lock_bh(&est_lock);
*pest = est->next;
write_unlock_bh(&est_lock);
kfree(est);
killed++;
}
if (killed && elist[idx].list == NULL)
del_timer(&elist[idx].timer);
}
}
EXPORT_SYMBOL(qdisc_kill_estimator);
EXPORT_SYMBOL(qdisc_new_estimator);

231
net/sched/gact.c Normal file
View File

@@ -0,0 +1,231 @@
/*
* net/sched/gact.c Generic actions
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* copyright Jamal Hadi Salim (2002-4)
*
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/config.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
#include <linux/tc_act/tc_gact.h>
#include <net/tc_act/tc_gact.h>
/* use generic hash table */
#define MY_TAB_SIZE 16
#define MY_TAB_MASK 15
static u32 idx_gen;
static struct tcf_gact *tcf_gact_ht[MY_TAB_SIZE];
static DEFINE_RWLOCK(gact_lock);
/* ovewrride the defaults */
#define tcf_st tcf_gact
#define tc_st tc_gact
#define tcf_t_lock gact_lock
#define tcf_ht tcf_gact_ht
#define CONFIG_NET_ACT_INIT 1
#include <net/pkt_act.h>
#ifdef CONFIG_GACT_PROB
static int gact_net_rand(struct tcf_gact *p)
{
if (net_random()%p->pval)
return p->action;
return p->paction;
}
static int gact_determ(struct tcf_gact *p)
{
if (p->bstats.packets%p->pval)
return p->action;
return p->paction;
}
typedef int (*g_rand)(struct tcf_gact *p);
static g_rand gact_rand[MAX_RAND]= { NULL, gact_net_rand, gact_determ };
#endif
static int tcf_gact_init(struct rtattr *rta, struct rtattr *est,
struct tc_action *a, int ovr, int bind)
{
struct rtattr *tb[TCA_GACT_MAX];
struct tc_gact *parm;
struct tcf_gact *p;
int ret = 0;
if (rta == NULL || rtattr_parse_nested(tb, TCA_GACT_MAX, rta) < 0)
return -EINVAL;
if (tb[TCA_GACT_PARMS - 1] == NULL ||
RTA_PAYLOAD(tb[TCA_GACT_PARMS - 1]) < sizeof(*parm))
return -EINVAL;
parm = RTA_DATA(tb[TCA_GACT_PARMS - 1]);
if (tb[TCA_GACT_PROB-1] != NULL)
#ifdef CONFIG_GACT_PROB
if (RTA_PAYLOAD(tb[TCA_GACT_PROB-1]) < sizeof(struct tc_gact_p))
return -EINVAL;
#else
return -EOPNOTSUPP;
#endif
p = tcf_hash_check(parm->index, a, ovr, bind);
if (p == NULL) {
p = tcf_hash_create(parm->index, est, a, sizeof(*p), ovr, bind);
if (p == NULL)
return -ENOMEM;
ret = ACT_P_CREATED;
} else {
if (!ovr) {
tcf_hash_release(p, bind);
return -EEXIST;
}
}
spin_lock_bh(&p->lock);
p->action = parm->action;
#ifdef CONFIG_GACT_PROB
if (tb[TCA_GACT_PROB-1] != NULL) {
struct tc_gact_p *p_parm = RTA_DATA(tb[TCA_GACT_PROB-1]);
p->paction = p_parm->paction;
p->pval = p_parm->pval;
p->ptype = p_parm->ptype;
}
#endif
spin_unlock_bh(&p->lock);
if (ret == ACT_P_CREATED)
tcf_hash_insert(p);
return ret;
}
static int
tcf_gact_cleanup(struct tc_action *a, int bind)
{
struct tcf_gact *p = PRIV(a, gact);
if (p != NULL)
return tcf_hash_release(p, bind);
return 0;
}
static int
tcf_gact(struct sk_buff **pskb, struct tc_action *a)
{
struct tcf_gact *p = PRIV(a, gact);
struct sk_buff *skb = *pskb;
int action = TC_ACT_SHOT;
spin_lock(&p->lock);
#ifdef CONFIG_GACT_PROB
if (p->ptype && gact_rand[p->ptype] != NULL)
action = gact_rand[p->ptype](p);
else
action = p->action;
#else
action = p->action;
#endif
p->bstats.bytes += skb->len;
p->bstats.packets++;
if (action == TC_ACT_SHOT)
p->qstats.drops++;
p->tm.lastuse = jiffies;
spin_unlock(&p->lock);
return action;
}
static int
tcf_gact_dump(struct sk_buff *skb, struct tc_action *a, int bind, int ref)
{
unsigned char *b = skb->tail;
struct tc_gact opt;
struct tcf_gact *p = PRIV(a, gact);
struct tcf_t t;
opt.index = p->index;
opt.refcnt = p->refcnt - ref;
opt.bindcnt = p->bindcnt - bind;
opt.action = p->action;
RTA_PUT(skb, TCA_GACT_PARMS, sizeof(opt), &opt);
#ifdef CONFIG_GACT_PROB
if (p->ptype) {
struct tc_gact_p p_opt;
p_opt.paction = p->paction;
p_opt.pval = p->pval;
p_opt.ptype = p->ptype;
RTA_PUT(skb, TCA_GACT_PROB, sizeof(p_opt), &p_opt);
}
#endif
t.install = jiffies_to_clock_t(jiffies - p->tm.install);
t.lastuse = jiffies_to_clock_t(jiffies - p->tm.lastuse);
t.expires = jiffies_to_clock_t(p->tm.expires);
RTA_PUT(skb, TCA_GACT_TM, sizeof(t), &t);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tc_action_ops act_gact_ops = {
.kind = "gact",
.type = TCA_ACT_GACT,
.capab = TCA_CAP_NONE,
.owner = THIS_MODULE,
.act = tcf_gact,
.dump = tcf_gact_dump,
.cleanup = tcf_gact_cleanup,
.lookup = tcf_hash_search,
.init = tcf_gact_init,
.walk = tcf_generic_walker
};
MODULE_AUTHOR("Jamal Hadi Salim(2002-4)");
MODULE_DESCRIPTION("Generic Classifier actions");
MODULE_LICENSE("GPL");
static int __init
gact_init_module(void)
{
#ifdef CONFIG_GACT_PROB
printk("GACT probability on\n");
#else
printk("GACT probability NOT on\n");
#endif
return tcf_register_action(&act_gact_ops);
}
static void __exit
gact_cleanup_module(void)
{
tcf_unregister_action(&act_gact_ops);
}
module_init(gact_init_module);
module_exit(gact_cleanup_module);

326
net/sched/ipt.c Normal file
View File

@@ -0,0 +1,326 @@
/*
* net/sched/ipt.c iptables target interface
*
*TODO: Add other tables. For now we only support the ipv4 table targets
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Copyright: Jamal Hadi Salim (2002-4)
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <asm/bitops.h>
#include <linux/config.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <linux/kmod.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
#include <linux/tc_act/tc_ipt.h>
#include <net/tc_act/tc_ipt.h>
#include <linux/netfilter_ipv4/ip_tables.h>
/* use generic hash table */
#define MY_TAB_SIZE 16
#define MY_TAB_MASK 15
static u32 idx_gen;
static struct tcf_ipt *tcf_ipt_ht[MY_TAB_SIZE];
/* ipt hash table lock */
static DEFINE_RWLOCK(ipt_lock);
/* ovewrride the defaults */
#define tcf_st tcf_ipt
#define tcf_t_lock ipt_lock
#define tcf_ht tcf_ipt_ht
#define CONFIG_NET_ACT_INIT
#include <net/pkt_act.h>
static int
ipt_init_target(struct ipt_entry_target *t, char *table, unsigned int hook)
{
struct ipt_target *target;
int ret = 0;
target = ipt_find_target(t->u.user.name, t->u.user.revision);
if (!target)
return -ENOENT;
DPRINTK("ipt_init_target: found %s\n", target->name);
t->u.kernel.target = target;
if (t->u.kernel.target->checkentry
&& !t->u.kernel.target->checkentry(table, NULL, t->data,
t->u.target_size - sizeof(*t),
hook)) {
DPRINTK("ipt_init_target: check failed for `%s'.\n",
t->u.kernel.target->name);
module_put(t->u.kernel.target->me);
ret = -EINVAL;
}
return ret;
}
static void
ipt_destroy_target(struct ipt_entry_target *t)
{
if (t->u.kernel.target->destroy)
t->u.kernel.target->destroy(t->data,
t->u.target_size - sizeof(*t));
module_put(t->u.kernel.target->me);
}
static int
tcf_ipt_release(struct tcf_ipt *p, int bind)
{
int ret = 0;
if (p) {
if (bind)
p->bindcnt--;
p->refcnt--;
if (p->bindcnt <= 0 && p->refcnt <= 0) {
ipt_destroy_target(p->t);
kfree(p->tname);
kfree(p->t);
tcf_hash_destroy(p);
ret = ACT_P_DELETED;
}
}
return ret;
}
static int
tcf_ipt_init(struct rtattr *rta, struct rtattr *est, struct tc_action *a,
int ovr, int bind)
{
struct rtattr *tb[TCA_IPT_MAX];
struct tcf_ipt *p;
struct ipt_entry_target *td, *t;
char *tname;
int ret = 0, err;
u32 hook = 0;
u32 index = 0;
if (rta == NULL || rtattr_parse_nested(tb, TCA_IPT_MAX, rta) < 0)
return -EINVAL;
if (tb[TCA_IPT_HOOK-1] == NULL ||
RTA_PAYLOAD(tb[TCA_IPT_HOOK-1]) < sizeof(u32))
return -EINVAL;
if (tb[TCA_IPT_TARG-1] == NULL ||
RTA_PAYLOAD(tb[TCA_IPT_TARG-1]) < sizeof(*t))
return -EINVAL;
td = (struct ipt_entry_target *)RTA_DATA(tb[TCA_IPT_TARG-1]);
if (RTA_PAYLOAD(tb[TCA_IPT_TARG-1]) < td->u.target_size)
return -EINVAL;
if (tb[TCA_IPT_INDEX-1] != NULL &&
RTA_PAYLOAD(tb[TCA_IPT_INDEX-1]) >= sizeof(u32))
index = *(u32 *)RTA_DATA(tb[TCA_IPT_INDEX-1]);
p = tcf_hash_check(index, a, ovr, bind);
if (p == NULL) {
p = tcf_hash_create(index, est, a, sizeof(*p), ovr, bind);
if (p == NULL)
return -ENOMEM;
ret = ACT_P_CREATED;
} else {
if (!ovr) {
tcf_ipt_release(p, bind);
return -EEXIST;
}
}
hook = *(u32 *)RTA_DATA(tb[TCA_IPT_HOOK-1]);
err = -ENOMEM;
tname = kmalloc(IFNAMSIZ, GFP_KERNEL);
if (tname == NULL)
goto err1;
if (tb[TCA_IPT_TABLE - 1] == NULL ||
rtattr_strlcpy(tname, tb[TCA_IPT_TABLE-1], IFNAMSIZ) >= IFNAMSIZ)
strcpy(tname, "mangle");
t = kmalloc(td->u.target_size, GFP_KERNEL);
if (t == NULL)
goto err2;
memcpy(t, td, td->u.target_size);
if ((err = ipt_init_target(t, tname, hook)) < 0)
goto err3;
spin_lock_bh(&p->lock);
if (ret != ACT_P_CREATED) {
ipt_destroy_target(p->t);
kfree(p->tname);
kfree(p->t);
}
p->tname = tname;
p->t = t;
p->hook = hook;
spin_unlock_bh(&p->lock);
if (ret == ACT_P_CREATED)
tcf_hash_insert(p);
return ret;
err3:
kfree(t);
err2:
kfree(tname);
err1:
kfree(p);
return err;
}
static int
tcf_ipt_cleanup(struct tc_action *a, int bind)
{
struct tcf_ipt *p = PRIV(a, ipt);
return tcf_ipt_release(p, bind);
}
static int
tcf_ipt(struct sk_buff **pskb, struct tc_action *a)
{
int ret = 0, result = 0;
struct tcf_ipt *p = PRIV(a, ipt);
struct sk_buff *skb = *pskb;
if (skb_cloned(skb)) {
if (pskb_expand_head(skb, 0, 0, GFP_ATOMIC))
return TC_ACT_UNSPEC;
}
spin_lock(&p->lock);
p->tm.lastuse = jiffies;
p->bstats.bytes += skb->len;
p->bstats.packets++;
/* yes, we have to worry about both in and out dev
worry later - danger - this API seems to have changed
from earlier kernels */
ret = p->t->u.kernel.target->target(&skb, skb->dev, NULL,
p->hook, p->t->data, NULL);
switch (ret) {
case NF_ACCEPT:
result = TC_ACT_OK;
break;
case NF_DROP:
result = TC_ACT_SHOT;
p->qstats.drops++;
break;
case IPT_CONTINUE:
result = TC_ACT_PIPE;
break;
default:
if (net_ratelimit())
printk("Bogus netfilter code %d assume ACCEPT\n", ret);
result = TC_POLICE_OK;
break;
}
spin_unlock(&p->lock);
return result;
}
static int
tcf_ipt_dump(struct sk_buff *skb, struct tc_action *a, int bind, int ref)
{
struct ipt_entry_target *t;
struct tcf_t tm;
struct tc_cnt c;
unsigned char *b = skb->tail;
struct tcf_ipt *p = PRIV(a, ipt);
/* for simple targets kernel size == user size
** user name = target name
** for foolproof you need to not assume this
*/
t = kmalloc(p->t->u.user.target_size, GFP_ATOMIC);
if (t == NULL)
goto rtattr_failure;
c.bindcnt = p->bindcnt - bind;
c.refcnt = p->refcnt - ref;
memcpy(t, p->t, p->t->u.user.target_size);
strcpy(t->u.user.name, p->t->u.kernel.target->name);
DPRINTK("\ttcf_ipt_dump tablename %s length %d\n", p->tname,
strlen(p->tname));
DPRINTK("\tdump target name %s size %d size user %d "
"data[0] %x data[1] %x\n", p->t->u.kernel.target->name,
p->t->u.target_size, p->t->u.user.target_size,
p->t->data[0], p->t->data[1]);
RTA_PUT(skb, TCA_IPT_TARG, p->t->u.user.target_size, t);
RTA_PUT(skb, TCA_IPT_INDEX, 4, &p->index);
RTA_PUT(skb, TCA_IPT_HOOK, 4, &p->hook);
RTA_PUT(skb, TCA_IPT_CNT, sizeof(struct tc_cnt), &c);
RTA_PUT(skb, TCA_IPT_TABLE, IFNAMSIZ, p->tname);
tm.install = jiffies_to_clock_t(jiffies - p->tm.install);
tm.lastuse = jiffies_to_clock_t(jiffies - p->tm.lastuse);
tm.expires = jiffies_to_clock_t(p->tm.expires);
RTA_PUT(skb, TCA_IPT_TM, sizeof (tm), &tm);
kfree(t);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
kfree(t);
return -1;
}
static struct tc_action_ops act_ipt_ops = {
.kind = "ipt",
.type = TCA_ACT_IPT,
.capab = TCA_CAP_NONE,
.owner = THIS_MODULE,
.act = tcf_ipt,
.dump = tcf_ipt_dump,
.cleanup = tcf_ipt_cleanup,
.lookup = tcf_hash_search,
.init = tcf_ipt_init,
.walk = tcf_generic_walker
};
MODULE_AUTHOR("Jamal Hadi Salim(2002-4)");
MODULE_DESCRIPTION("Iptables target actions");
MODULE_LICENSE("GPL");
static int __init
ipt_init_module(void)
{
return tcf_register_action(&act_ipt_ops);
}
static void __exit
ipt_cleanup_module(void)
{
tcf_unregister_action(&act_ipt_ops);
}
module_init(ipt_init_module);
module_exit(ipt_cleanup_module);

276
net/sched/mirred.c Normal file
View File

@@ -0,0 +1,276 @@
/*
* net/sched/mirred.c packet mirroring and redirect actions
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Jamal Hadi Salim (2002-4)
*
* TODO: Add ingress support (and socket redirect support)
*
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <asm/bitops.h>
#include <linux/config.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
#include <linux/tc_act/tc_mirred.h>
#include <net/tc_act/tc_mirred.h>
#include <linux/etherdevice.h>
#include <linux/if_arp.h>
/* use generic hash table */
#define MY_TAB_SIZE 8
#define MY_TAB_MASK (MY_TAB_SIZE - 1)
static u32 idx_gen;
static struct tcf_mirred *tcf_mirred_ht[MY_TAB_SIZE];
static DEFINE_RWLOCK(mirred_lock);
/* ovewrride the defaults */
#define tcf_st tcf_mirred
#define tc_st tc_mirred
#define tcf_t_lock mirred_lock
#define tcf_ht tcf_mirred_ht
#define CONFIG_NET_ACT_INIT 1
#include <net/pkt_act.h>
static inline int
tcf_mirred_release(struct tcf_mirred *p, int bind)
{
if (p) {
if (bind)
p->bindcnt--;
p->refcnt--;
if(!p->bindcnt && p->refcnt <= 0) {
dev_put(p->dev);
tcf_hash_destroy(p);
return 1;
}
}
return 0;
}
static int
tcf_mirred_init(struct rtattr *rta, struct rtattr *est, struct tc_action *a,
int ovr, int bind)
{
struct rtattr *tb[TCA_MIRRED_MAX];
struct tc_mirred *parm;
struct tcf_mirred *p;
struct net_device *dev = NULL;
int ret = 0;
int ok_push = 0;
if (rta == NULL || rtattr_parse_nested(tb, TCA_MIRRED_MAX, rta) < 0)
return -EINVAL;
if (tb[TCA_MIRRED_PARMS-1] == NULL ||
RTA_PAYLOAD(tb[TCA_MIRRED_PARMS-1]) < sizeof(*parm))
return -EINVAL;
parm = RTA_DATA(tb[TCA_MIRRED_PARMS-1]);
if (parm->ifindex) {
dev = __dev_get_by_index(parm->ifindex);
if (dev == NULL)
return -ENODEV;
switch (dev->type) {
case ARPHRD_TUNNEL:
case ARPHRD_TUNNEL6:
case ARPHRD_SIT:
case ARPHRD_IPGRE:
case ARPHRD_VOID:
case ARPHRD_NONE:
ok_push = 0;
break;
default:
ok_push = 1;
break;
}
}
p = tcf_hash_check(parm->index, a, ovr, bind);
if (p == NULL) {
if (!parm->ifindex)
return -EINVAL;
p = tcf_hash_create(parm->index, est, a, sizeof(*p), ovr, bind);
if (p == NULL)
return -ENOMEM;
ret = ACT_P_CREATED;
} else {
if (!ovr) {
tcf_mirred_release(p, bind);
return -EEXIST;
}
}
spin_lock_bh(&p->lock);
p->action = parm->action;
p->eaction = parm->eaction;
if (parm->ifindex) {
p->ifindex = parm->ifindex;
if (ret != ACT_P_CREATED)
dev_put(p->dev);
p->dev = dev;
dev_hold(dev);
p->ok_push = ok_push;
}
spin_unlock_bh(&p->lock);
if (ret == ACT_P_CREATED)
tcf_hash_insert(p);
DPRINTK("tcf_mirred_init index %d action %d eaction %d device %s "
"ifindex %d\n", parm->index, parm->action, parm->eaction,
dev->name, parm->ifindex);
return ret;
}
static int
tcf_mirred_cleanup(struct tc_action *a, int bind)
{
struct tcf_mirred *p = PRIV(a, mirred);
if (p != NULL)
return tcf_mirred_release(p, bind);
return 0;
}
static int
tcf_mirred(struct sk_buff **pskb, struct tc_action *a)
{
struct tcf_mirred *p = PRIV(a, mirred);
struct net_device *dev;
struct sk_buff *skb2 = NULL;
struct sk_buff *skb = *pskb;
u32 at = G_TC_AT(skb->tc_verd);
spin_lock(&p->lock);
dev = p->dev;
p->tm.lastuse = jiffies;
if (!(dev->flags&IFF_UP) ) {
if (net_ratelimit())
printk("mirred to Houston: device %s is gone!\n",
dev->name);
bad_mirred:
if (skb2 != NULL)
kfree_skb(skb2);
p->qstats.overlimits++;
p->bstats.bytes += skb->len;
p->bstats.packets++;
spin_unlock(&p->lock);
/* should we be asking for packet to be dropped?
* may make sense for redirect case only
*/
return TC_ACT_SHOT;
}
skb2 = skb_clone(skb, GFP_ATOMIC);
if (skb2 == NULL)
goto bad_mirred;
if (p->eaction != TCA_EGRESS_MIRROR && p->eaction != TCA_EGRESS_REDIR) {
if (net_ratelimit())
printk("tcf_mirred unknown action %d\n", p->eaction);
goto bad_mirred;
}
p->bstats.bytes += skb2->len;
p->bstats.packets++;
if (!(at & AT_EGRESS))
if (p->ok_push)
skb_push(skb2, skb2->dev->hard_header_len);
/* mirror is always swallowed */
if (p->eaction != TCA_EGRESS_MIRROR)
skb2->tc_verd = SET_TC_FROM(skb2->tc_verd, at);
skb2->dev = dev;
skb2->input_dev = skb->dev;
dev_queue_xmit(skb2);
spin_unlock(&p->lock);
return p->action;
}
static int
tcf_mirred_dump(struct sk_buff *skb, struct tc_action *a, int bind, int ref)
{
unsigned char *b = skb->tail;
struct tc_mirred opt;
struct tcf_mirred *p = PRIV(a, mirred);
struct tcf_t t;
opt.index = p->index;
opt.action = p->action;
opt.refcnt = p->refcnt - ref;
opt.bindcnt = p->bindcnt - bind;
opt.eaction = p->eaction;
opt.ifindex = p->ifindex;
DPRINTK("tcf_mirred_dump index %d action %d eaction %d ifindex %d\n",
p->index, p->action, p->eaction, p->ifindex);
RTA_PUT(skb, TCA_MIRRED_PARMS, sizeof(opt), &opt);
t.install = jiffies_to_clock_t(jiffies - p->tm.install);
t.lastuse = jiffies_to_clock_t(jiffies - p->tm.lastuse);
t.expires = jiffies_to_clock_t(p->tm.expires);
RTA_PUT(skb, TCA_MIRRED_TM, sizeof(t), &t);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct tc_action_ops act_mirred_ops = {
.kind = "mirred",
.type = TCA_ACT_MIRRED,
.capab = TCA_CAP_NONE,
.owner = THIS_MODULE,
.act = tcf_mirred,
.dump = tcf_mirred_dump,
.cleanup = tcf_mirred_cleanup,
.lookup = tcf_hash_search,
.init = tcf_mirred_init,
.walk = tcf_generic_walker
};
MODULE_AUTHOR("Jamal Hadi Salim(2002)");
MODULE_DESCRIPTION("Device Mirror/redirect actions");
MODULE_LICENSE("GPL");
static int __init
mirred_init_module(void)
{
printk("Mirror/redirect action on\n");
return tcf_register_action(&act_mirred_ops);
}
static void __exit
mirred_cleanup_module(void)
{
tcf_unregister_action(&act_mirred_ops);
}
module_init(mirred_init_module);
module_exit(mirred_cleanup_module);

288
net/sched/pedit.c Normal file
View File

@@ -0,0 +1,288 @@
/*
* net/sched/pedit.c Generic packet editor
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Jamal Hadi Salim (2002-4)
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <asm/bitops.h>
#include <linux/config.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
#include <linux/tc_act/tc_pedit.h>
#include <net/tc_act/tc_pedit.h>
#define PEDIT_DEB 1
/* use generic hash table */
#define MY_TAB_SIZE 16
#define MY_TAB_MASK 15
static u32 idx_gen;
static struct tcf_pedit *tcf_pedit_ht[MY_TAB_SIZE];
static DEFINE_RWLOCK(pedit_lock);
#define tcf_st tcf_pedit
#define tc_st tc_pedit
#define tcf_t_lock pedit_lock
#define tcf_ht tcf_pedit_ht
#define CONFIG_NET_ACT_INIT 1
#include <net/pkt_act.h>
static int
tcf_pedit_init(struct rtattr *rta, struct rtattr *est, struct tc_action *a,
int ovr, int bind)
{
struct rtattr *tb[TCA_PEDIT_MAX];
struct tc_pedit *parm;
int ret = 0;
struct tcf_pedit *p;
struct tc_pedit_key *keys = NULL;
int ksize;
if (rta == NULL || rtattr_parse_nested(tb, TCA_PEDIT_MAX, rta) < 0)
return -EINVAL;
if (tb[TCA_PEDIT_PARMS - 1] == NULL ||
RTA_PAYLOAD(tb[TCA_PEDIT_PARMS-1]) < sizeof(*parm))
return -EINVAL;
parm = RTA_DATA(tb[TCA_PEDIT_PARMS-1]);
ksize = parm->nkeys * sizeof(struct tc_pedit_key);
if (RTA_PAYLOAD(tb[TCA_PEDIT_PARMS-1]) < sizeof(*parm) + ksize)
return -EINVAL;
p = tcf_hash_check(parm->index, a, ovr, bind);
if (p == NULL) {
if (!parm->nkeys)
return -EINVAL;
p = tcf_hash_create(parm->index, est, a, sizeof(*p), ovr, bind);
if (p == NULL)
return -ENOMEM;
keys = kmalloc(ksize, GFP_KERNEL);
if (keys == NULL) {
kfree(p);
return -ENOMEM;
}
ret = ACT_P_CREATED;
} else {
if (!ovr) {
tcf_hash_release(p, bind);
return -EEXIST;
}
if (p->nkeys && p->nkeys != parm->nkeys) {
keys = kmalloc(ksize, GFP_KERNEL);
if (keys == NULL)
return -ENOMEM;
}
}
spin_lock_bh(&p->lock);
p->flags = parm->flags;
p->action = parm->action;
if (keys) {
kfree(p->keys);
p->keys = keys;
p->nkeys = parm->nkeys;
}
memcpy(p->keys, parm->keys, ksize);
spin_unlock_bh(&p->lock);
if (ret == ACT_P_CREATED)
tcf_hash_insert(p);
return ret;
}
static int
tcf_pedit_cleanup(struct tc_action *a, int bind)
{
struct tcf_pedit *p = PRIV(a, pedit);
if (p != NULL) {
struct tc_pedit_key *keys = p->keys;
if (tcf_hash_release(p, bind)) {
kfree(keys);
return 1;
}
}
return 0;
}
static int
tcf_pedit(struct sk_buff **pskb, struct tc_action *a)
{
struct tcf_pedit *p = PRIV(a, pedit);
struct sk_buff *skb = *pskb;
int i, munged = 0;
u8 *pptr;
if (!(skb->tc_verd & TC_OK2MUNGE)) {
/* should we set skb->cloned? */
if (pskb_expand_head(skb, 0, 0, GFP_ATOMIC)) {
return p->action;
}
}
pptr = skb->nh.raw;
spin_lock(&p->lock);
p->tm.lastuse = jiffies;
if (p->nkeys > 0) {
struct tc_pedit_key *tkey = p->keys;
for (i = p->nkeys; i > 0; i--, tkey++) {
u32 *ptr;
int offset = tkey->off;
if (tkey->offmask) {
if (skb->len > tkey->at) {
char *j = pptr + tkey->at;
offset += ((*j & tkey->offmask) >>
tkey->shift);
} else {
goto bad;
}
}
if (offset % 4) {
printk("offset must be on 32 bit boundaries\n");
goto bad;
}
if (skb->len < 0 || (offset > 0 && offset > skb->len)) {
printk("offset %d cant exceed pkt length %d\n",
offset, skb->len);
goto bad;
}
ptr = (u32 *)(pptr+offset);
/* just do it, baby */
*ptr = ((*ptr & tkey->mask) ^ tkey->val);
munged++;
}
if (munged)
skb->tc_verd = SET_TC_MUNGED(skb->tc_verd);
goto done;
} else {
printk("pedit BUG: index %d\n",p->index);
}
bad:
p->qstats.overlimits++;
done:
p->bstats.bytes += skb->len;
p->bstats.packets++;
spin_unlock(&p->lock);
return p->action;
}
static int
tcf_pedit_dump(struct sk_buff *skb, struct tc_action *a,int bind, int ref)
{
unsigned char *b = skb->tail;
struct tc_pedit *opt;
struct tcf_pedit *p = PRIV(a, pedit);
struct tcf_t t;
int s;
s = sizeof(*opt) + p->nkeys * sizeof(struct tc_pedit_key);
/* netlink spinlocks held above us - must use ATOMIC */
opt = kmalloc(s, GFP_ATOMIC);
if (opt == NULL)
return -ENOBUFS;
memset(opt, 0, s);
memcpy(opt->keys, p->keys, p->nkeys * sizeof(struct tc_pedit_key));
opt->index = p->index;
opt->nkeys = p->nkeys;
opt->flags = p->flags;
opt->action = p->action;
opt->refcnt = p->refcnt - ref;
opt->bindcnt = p->bindcnt - bind;
#ifdef PEDIT_DEB
{
/* Debug - get rid of later */
int i;
struct tc_pedit_key *key = opt->keys;
for (i=0; i<opt->nkeys; i++, key++) {
printk( "\n key #%d",i);
printk( " at %d: val %08x mask %08x",
(unsigned int)key->off,
(unsigned int)key->val,
(unsigned int)key->mask);
}
}
#endif
RTA_PUT(skb, TCA_PEDIT_PARMS, s, opt);
t.install = jiffies_to_clock_t(jiffies - p->tm.install);
t.lastuse = jiffies_to_clock_t(jiffies - p->tm.lastuse);
t.expires = jiffies_to_clock_t(p->tm.expires);
RTA_PUT(skb, TCA_PEDIT_TM, sizeof(t), &t);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static
struct tc_action_ops act_pedit_ops = {
.kind = "pedit",
.type = TCA_ACT_PEDIT,
.capab = TCA_CAP_NONE,
.owner = THIS_MODULE,
.act = tcf_pedit,
.dump = tcf_pedit_dump,
.cleanup = tcf_pedit_cleanup,
.lookup = tcf_hash_search,
.init = tcf_pedit_init,
.walk = tcf_generic_walker
};
MODULE_AUTHOR("Jamal Hadi Salim(2002-4)");
MODULE_DESCRIPTION("Generic Packet Editor actions");
MODULE_LICENSE("GPL");
static int __init
pedit_init_module(void)
{
return tcf_register_action(&act_pedit_ops);
}
static void __exit
pedit_cleanup_module(void)
{
tcf_unregister_action(&act_pedit_ops);
}
module_init(pedit_init_module);
module_exit(pedit_cleanup_module);

612
net/sched/police.c Normal file
View File

@@ -0,0 +1,612 @@
/*
* net/sched/police.c Input police filter.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
* J Hadi Salim (action changes)
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/module.h>
#include <linux/rtnetlink.h>
#include <linux/init.h>
#include <net/sock.h>
#include <net/act_api.h>
#define L2T(p,L) ((p)->R_tab->data[(L)>>(p)->R_tab->rate.cell_log])
#define L2T_P(p,L) ((p)->P_tab->data[(L)>>(p)->P_tab->rate.cell_log])
#define PRIV(a) ((struct tcf_police *) (a)->priv)
/* use generic hash table */
#define MY_TAB_SIZE 16
#define MY_TAB_MASK 15
static u32 idx_gen;
static struct tcf_police *tcf_police_ht[MY_TAB_SIZE];
/* Policer hash table lock */
static DEFINE_RWLOCK(police_lock);
/* Each policer is serialized by its individual spinlock */
static __inline__ unsigned tcf_police_hash(u32 index)
{
return index&0xF;
}
static __inline__ struct tcf_police * tcf_police_lookup(u32 index)
{
struct tcf_police *p;
read_lock(&police_lock);
for (p = tcf_police_ht[tcf_police_hash(index)]; p; p = p->next) {
if (p->index == index)
break;
}
read_unlock(&police_lock);
return p;
}
#ifdef CONFIG_NET_CLS_ACT
static int tcf_generic_walker(struct sk_buff *skb, struct netlink_callback *cb,
int type, struct tc_action *a)
{
struct tcf_police *p;
int err = 0, index = -1, i = 0, s_i = 0, n_i = 0;
struct rtattr *r;
read_lock(&police_lock);
s_i = cb->args[0];
for (i = 0; i < MY_TAB_SIZE; i++) {
p = tcf_police_ht[tcf_police_hash(i)];
for (; p; p = p->next) {
index++;
if (index < s_i)
continue;
a->priv = p;
a->order = index;
r = (struct rtattr*) skb->tail;
RTA_PUT(skb, a->order, 0, NULL);
if (type == RTM_DELACTION)
err = tcf_action_dump_1(skb, a, 0, 1);
else
err = tcf_action_dump_1(skb, a, 0, 0);
if (err < 0) {
index--;
skb_trim(skb, (u8*)r - skb->data);
goto done;
}
r->rta_len = skb->tail - (u8*)r;
n_i++;
}
}
done:
read_unlock(&police_lock);
if (n_i)
cb->args[0] += n_i;
return n_i;
rtattr_failure:
skb_trim(skb, (u8*)r - skb->data);
goto done;
}
static inline int
tcf_hash_search(struct tc_action *a, u32 index)
{
struct tcf_police *p = tcf_police_lookup(index);
if (p != NULL) {
a->priv = p;
return 1;
} else {
return 0;
}
}
#endif
static inline u32 tcf_police_new_index(void)
{
do {
if (++idx_gen == 0)
idx_gen = 1;
} while (tcf_police_lookup(idx_gen));
return idx_gen;
}
void tcf_police_destroy(struct tcf_police *p)
{
unsigned h = tcf_police_hash(p->index);
struct tcf_police **p1p;
for (p1p = &tcf_police_ht[h]; *p1p; p1p = &(*p1p)->next) {
if (*p1p == p) {
write_lock_bh(&police_lock);
*p1p = p->next;
write_unlock_bh(&police_lock);
#ifdef CONFIG_NET_ESTIMATOR
gen_kill_estimator(&p->bstats, &p->rate_est);
#endif
if (p->R_tab)
qdisc_put_rtab(p->R_tab);
if (p->P_tab)
qdisc_put_rtab(p->P_tab);
kfree(p);
return;
}
}
BUG_TRAP(0);
}
#ifdef CONFIG_NET_CLS_ACT
static int tcf_act_police_locate(struct rtattr *rta, struct rtattr *est,
struct tc_action *a, int ovr, int bind)
{
unsigned h;
int ret = 0, err;
struct rtattr *tb[TCA_POLICE_MAX];
struct tc_police *parm;
struct tcf_police *p;
struct qdisc_rate_table *R_tab = NULL, *P_tab = NULL;
if (rta == NULL || rtattr_parse_nested(tb, TCA_POLICE_MAX, rta) < 0)
return -EINVAL;
if (tb[TCA_POLICE_TBF-1] == NULL ||
RTA_PAYLOAD(tb[TCA_POLICE_TBF-1]) != sizeof(*parm))
return -EINVAL;
parm = RTA_DATA(tb[TCA_POLICE_TBF-1]);
if (tb[TCA_POLICE_RESULT-1] != NULL &&
RTA_PAYLOAD(tb[TCA_POLICE_RESULT-1]) != sizeof(u32))
return -EINVAL;
if (tb[TCA_POLICE_RESULT-1] != NULL &&
RTA_PAYLOAD(tb[TCA_POLICE_RESULT-1]) != sizeof(u32))
return -EINVAL;
if (parm->index && (p = tcf_police_lookup(parm->index)) != NULL) {
a->priv = p;
if (bind) {
p->bindcnt += 1;
p->refcnt += 1;
}
if (ovr)
goto override;
return ret;
}
p = kmalloc(sizeof(*p), GFP_KERNEL);
if (p == NULL)
return -ENOMEM;
memset(p, 0, sizeof(*p));
ret = ACT_P_CREATED;
p->refcnt = 1;
spin_lock_init(&p->lock);
p->stats_lock = &p->lock;
if (bind)
p->bindcnt = 1;
override:
if (parm->rate.rate) {
err = -ENOMEM;
R_tab = qdisc_get_rtab(&parm->rate, tb[TCA_POLICE_RATE-1]);
if (R_tab == NULL)
goto failure;
if (parm->peakrate.rate) {
P_tab = qdisc_get_rtab(&parm->peakrate,
tb[TCA_POLICE_PEAKRATE-1]);
if (p->P_tab == NULL) {
qdisc_put_rtab(R_tab);
goto failure;
}
}
}
/* No failure allowed after this point */
spin_lock_bh(&p->lock);
if (R_tab != NULL) {
qdisc_put_rtab(p->R_tab);
p->R_tab = R_tab;
}
if (P_tab != NULL) {
qdisc_put_rtab(p->P_tab);
p->P_tab = P_tab;
}
if (tb[TCA_POLICE_RESULT-1])
p->result = *(u32*)RTA_DATA(tb[TCA_POLICE_RESULT-1]);
p->toks = p->burst = parm->burst;
p->mtu = parm->mtu;
if (p->mtu == 0) {
p->mtu = ~0;
if (p->R_tab)
p->mtu = 255<<p->R_tab->rate.cell_log;
}
if (p->P_tab)
p->ptoks = L2T_P(p, p->mtu);
p->action = parm->action;
#ifdef CONFIG_NET_ESTIMATOR
if (tb[TCA_POLICE_AVRATE-1])
p->ewma_rate = *(u32*)RTA_DATA(tb[TCA_POLICE_AVRATE-1]);
if (est)
gen_replace_estimator(&p->bstats, &p->rate_est, p->stats_lock, est);
#endif
spin_unlock_bh(&p->lock);
if (ret != ACT_P_CREATED)
return ret;
PSCHED_GET_TIME(p->t_c);
p->index = parm->index ? : tcf_police_new_index();
h = tcf_police_hash(p->index);
write_lock_bh(&police_lock);
p->next = tcf_police_ht[h];
tcf_police_ht[h] = p;
write_unlock_bh(&police_lock);
a->priv = p;
return ret;
failure:
if (ret == ACT_P_CREATED)
kfree(p);
return err;
}
static int tcf_act_police_cleanup(struct tc_action *a, int bind)
{
struct tcf_police *p = PRIV(a);
if (p != NULL)
return tcf_police_release(p, bind);
return 0;
}
static int tcf_act_police(struct sk_buff **pskb, struct tc_action *a)
{
psched_time_t now;
struct sk_buff *skb = *pskb;
struct tcf_police *p = PRIV(a);
long toks;
long ptoks = 0;
spin_lock(&p->lock);
p->bstats.bytes += skb->len;
p->bstats.packets++;
#ifdef CONFIG_NET_ESTIMATOR
if (p->ewma_rate && p->rate_est.bps >= p->ewma_rate) {
p->qstats.overlimits++;
spin_unlock(&p->lock);
return p->action;
}
#endif
if (skb->len <= p->mtu) {
if (p->R_tab == NULL) {
spin_unlock(&p->lock);
return p->result;
}
PSCHED_GET_TIME(now);
toks = PSCHED_TDIFF_SAFE(now, p->t_c, p->burst);
if (p->P_tab) {
ptoks = toks + p->ptoks;
if (ptoks > (long)L2T_P(p, p->mtu))
ptoks = (long)L2T_P(p, p->mtu);
ptoks -= L2T_P(p, skb->len);
}
toks += p->toks;
if (toks > (long)p->burst)
toks = p->burst;
toks -= L2T(p, skb->len);
if ((toks|ptoks) >= 0) {
p->t_c = now;
p->toks = toks;
p->ptoks = ptoks;
spin_unlock(&p->lock);
return p->result;
}
}
p->qstats.overlimits++;
spin_unlock(&p->lock);
return p->action;
}
static int
tcf_act_police_dump(struct sk_buff *skb, struct tc_action *a, int bind, int ref)
{
unsigned char *b = skb->tail;
struct tc_police opt;
struct tcf_police *p = PRIV(a);
opt.index = p->index;
opt.action = p->action;
opt.mtu = p->mtu;
opt.burst = p->burst;
opt.refcnt = p->refcnt - ref;
opt.bindcnt = p->bindcnt - bind;
if (p->R_tab)
opt.rate = p->R_tab->rate;
else
memset(&opt.rate, 0, sizeof(opt.rate));
if (p->P_tab)
opt.peakrate = p->P_tab->rate;
else
memset(&opt.peakrate, 0, sizeof(opt.peakrate));
RTA_PUT(skb, TCA_POLICE_TBF, sizeof(opt), &opt);
if (p->result)
RTA_PUT(skb, TCA_POLICE_RESULT, sizeof(int), &p->result);
#ifdef CONFIG_NET_ESTIMATOR
if (p->ewma_rate)
RTA_PUT(skb, TCA_POLICE_AVRATE, 4, &p->ewma_rate);
#endif
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
MODULE_AUTHOR("Alexey Kuznetsov");
MODULE_DESCRIPTION("Policing actions");
MODULE_LICENSE("GPL");
static struct tc_action_ops act_police_ops = {
.kind = "police",
.type = TCA_ID_POLICE,
.capab = TCA_CAP_NONE,
.owner = THIS_MODULE,
.act = tcf_act_police,
.dump = tcf_act_police_dump,
.cleanup = tcf_act_police_cleanup,
.lookup = tcf_hash_search,
.init = tcf_act_police_locate,
.walk = tcf_generic_walker
};
static int __init
police_init_module(void)
{
return tcf_register_action(&act_police_ops);
}
static void __exit
police_cleanup_module(void)
{
tcf_unregister_action(&act_police_ops);
}
module_init(police_init_module);
module_exit(police_cleanup_module);
#endif
struct tcf_police * tcf_police_locate(struct rtattr *rta, struct rtattr *est)
{
unsigned h;
struct tcf_police *p;
struct rtattr *tb[TCA_POLICE_MAX];
struct tc_police *parm;
if (rtattr_parse_nested(tb, TCA_POLICE_MAX, rta) < 0)
return NULL;
if (tb[TCA_POLICE_TBF-1] == NULL ||
RTA_PAYLOAD(tb[TCA_POLICE_TBF-1]) != sizeof(*parm))
return NULL;
parm = RTA_DATA(tb[TCA_POLICE_TBF-1]);
if (parm->index && (p = tcf_police_lookup(parm->index)) != NULL) {
p->refcnt++;
return p;
}
p = kmalloc(sizeof(*p), GFP_KERNEL);
if (p == NULL)
return NULL;
memset(p, 0, sizeof(*p));
p->refcnt = 1;
spin_lock_init(&p->lock);
p->stats_lock = &p->lock;
if (parm->rate.rate) {
p->R_tab = qdisc_get_rtab(&parm->rate, tb[TCA_POLICE_RATE-1]);
if (p->R_tab == NULL)
goto failure;
if (parm->peakrate.rate) {
p->P_tab = qdisc_get_rtab(&parm->peakrate,
tb[TCA_POLICE_PEAKRATE-1]);
if (p->P_tab == NULL)
goto failure;
}
}
if (tb[TCA_POLICE_RESULT-1]) {
if (RTA_PAYLOAD(tb[TCA_POLICE_RESULT-1]) != sizeof(u32))
goto failure;
p->result = *(u32*)RTA_DATA(tb[TCA_POLICE_RESULT-1]);
}
#ifdef CONFIG_NET_ESTIMATOR
if (tb[TCA_POLICE_AVRATE-1]) {
if (RTA_PAYLOAD(tb[TCA_POLICE_AVRATE-1]) != sizeof(u32))
goto failure;
p->ewma_rate = *(u32*)RTA_DATA(tb[TCA_POLICE_AVRATE-1]);
}
#endif
p->toks = p->burst = parm->burst;
p->mtu = parm->mtu;
if (p->mtu == 0) {
p->mtu = ~0;
if (p->R_tab)
p->mtu = 255<<p->R_tab->rate.cell_log;
}
if (p->P_tab)
p->ptoks = L2T_P(p, p->mtu);
PSCHED_GET_TIME(p->t_c);
p->index = parm->index ? : tcf_police_new_index();
p->action = parm->action;
#ifdef CONFIG_NET_ESTIMATOR
if (est)
gen_new_estimator(&p->bstats, &p->rate_est, p->stats_lock, est);
#endif
h = tcf_police_hash(p->index);
write_lock_bh(&police_lock);
p->next = tcf_police_ht[h];
tcf_police_ht[h] = p;
write_unlock_bh(&police_lock);
return p;
failure:
if (p->R_tab)
qdisc_put_rtab(p->R_tab);
kfree(p);
return NULL;
}
int tcf_police(struct sk_buff *skb, struct tcf_police *p)
{
psched_time_t now;
long toks;
long ptoks = 0;
spin_lock(&p->lock);
p->bstats.bytes += skb->len;
p->bstats.packets++;
#ifdef CONFIG_NET_ESTIMATOR
if (p->ewma_rate && p->rate_est.bps >= p->ewma_rate) {
p->qstats.overlimits++;
spin_unlock(&p->lock);
return p->action;
}
#endif
if (skb->len <= p->mtu) {
if (p->R_tab == NULL) {
spin_unlock(&p->lock);
return p->result;
}
PSCHED_GET_TIME(now);
toks = PSCHED_TDIFF_SAFE(now, p->t_c, p->burst);
if (p->P_tab) {
ptoks = toks + p->ptoks;
if (ptoks > (long)L2T_P(p, p->mtu))
ptoks = (long)L2T_P(p, p->mtu);
ptoks -= L2T_P(p, skb->len);
}
toks += p->toks;
if (toks > (long)p->burst)
toks = p->burst;
toks -= L2T(p, skb->len);
if ((toks|ptoks) >= 0) {
p->t_c = now;
p->toks = toks;
p->ptoks = ptoks;
spin_unlock(&p->lock);
return p->result;
}
}
p->qstats.overlimits++;
spin_unlock(&p->lock);
return p->action;
}
int tcf_police_dump(struct sk_buff *skb, struct tcf_police *p)
{
unsigned char *b = skb->tail;
struct tc_police opt;
opt.index = p->index;
opt.action = p->action;
opt.mtu = p->mtu;
opt.burst = p->burst;
if (p->R_tab)
opt.rate = p->R_tab->rate;
else
memset(&opt.rate, 0, sizeof(opt.rate));
if (p->P_tab)
opt.peakrate = p->P_tab->rate;
else
memset(&opt.peakrate, 0, sizeof(opt.peakrate));
RTA_PUT(skb, TCA_POLICE_TBF, sizeof(opt), &opt);
if (p->result)
RTA_PUT(skb, TCA_POLICE_RESULT, sizeof(int), &p->result);
#ifdef CONFIG_NET_ESTIMATOR
if (p->ewma_rate)
RTA_PUT(skb, TCA_POLICE_AVRATE, 4, &p->ewma_rate);
#endif
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
int tcf_police_dump_stats(struct sk_buff *skb, struct tcf_police *p)
{
struct gnet_dump d;
if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS,
TCA_XSTATS, p->stats_lock, &d) < 0)
goto errout;
if (gnet_stats_copy_basic(&d, &p->bstats) < 0 ||
#ifdef CONFIG_NET_ESTIMATOR
gnet_stats_copy_rate_est(&d, &p->rate_est) < 0 ||
#endif
gnet_stats_copy_queue(&d, &p->qstats) < 0)
goto errout;
if (gnet_stats_finish_copy(&d) < 0)
goto errout;
return 0;
errout:
return -1;
}
EXPORT_SYMBOL(tcf_police);
EXPORT_SYMBOL(tcf_police_destroy);
EXPORT_SYMBOL(tcf_police_dump);
EXPORT_SYMBOL(tcf_police_dump_stats);
EXPORT_SYMBOL(tcf_police_hash);
EXPORT_SYMBOL(tcf_police_ht);
EXPORT_SYMBOL(tcf_police_locate);
EXPORT_SYMBOL(tcf_police_lookup);
EXPORT_SYMBOL(tcf_police_new_index);

1296
net/sched/sch_api.c Normal file
View File

File diff suppressed because it is too large Load Diff

735
net/sched/sch_atm.c Normal file
View File

@@ -0,0 +1,735 @@
/* net/sched/sch_atm.c - ATM VC selection "queueing discipline" */
/* Written 1998-2000 by Werner Almesberger, EPFL ICA */
#include <linux/config.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/skbuff.h>
#include <linux/interrupt.h>
#include <linux/atmdev.h>
#include <linux/atmclip.h>
#include <linux/netdevice.h>
#include <linux/rtnetlink.h>
#include <linux/file.h> /* for fput */
#include <net/pkt_sched.h>
#include <net/sock.h>
extern struct socket *sockfd_lookup(int fd, int *err); /* @@@ fix this */
#if 0 /* control */
#define DPRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define DPRINTK(format,args...)
#endif
#if 0 /* data */
#define D2PRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define D2PRINTK(format,args...)
#endif
/*
* The ATM queuing discipline provides a framework for invoking classifiers
* (aka "filters"), which in turn select classes of this queuing discipline.
* Each class maps the flow(s) it is handling to a given VC. Multiple classes
* may share the same VC.
*
* When creating a class, VCs are specified by passing the number of the open
* socket descriptor by which the calling process references the VC. The kernel
* keeps the VC open at least until all classes using it are removed.
*
* In this file, most functions are named atm_tc_* to avoid confusion with all
* the atm_* in net/atm. This naming convention differs from what's used in the
* rest of net/sched.
*
* Known bugs:
* - sometimes messes up the IP stack
* - any manipulations besides the few operations described in the README, are
* untested and likely to crash the system
* - should lock the flow while there is data in the queue (?)
*/
#define PRIV(sch) qdisc_priv(sch)
#define VCC2FLOW(vcc) ((struct atm_flow_data *) ((vcc)->user_back))
struct atm_flow_data {
struct Qdisc *q; /* FIFO, TBF, etc. */
struct tcf_proto *filter_list;
struct atm_vcc *vcc; /* VCC; NULL if VCC is closed */
void (*old_pop)(struct atm_vcc *vcc,struct sk_buff *skb); /* chaining */
struct atm_qdisc_data *parent; /* parent qdisc */
struct socket *sock; /* for closing */
u32 classid; /* x:y type ID */
int ref; /* reference count */
struct gnet_stats_basic bstats;
struct gnet_stats_queue qstats;
spinlock_t *stats_lock;
struct atm_flow_data *next;
struct atm_flow_data *excess; /* flow for excess traffic;
NULL to set CLP instead */
int hdr_len;
unsigned char hdr[0]; /* header data; MUST BE LAST */
};
struct atm_qdisc_data {
struct atm_flow_data link; /* unclassified skbs go here */
struct atm_flow_data *flows; /* NB: "link" is also on this
list */
struct tasklet_struct task; /* requeue tasklet */
};
/* ------------------------- Class/flow operations ------------------------- */
static int find_flow(struct atm_qdisc_data *qdisc,struct atm_flow_data *flow)
{
struct atm_flow_data *walk;
DPRINTK("find_flow(qdisc %p,flow %p)\n",qdisc,flow);
for (walk = qdisc->flows; walk; walk = walk->next)
if (walk == flow) return 1;
DPRINTK("find_flow: not found\n");
return 0;
}
static __inline__ struct atm_flow_data *lookup_flow(struct Qdisc *sch,
u32 classid)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow;
for (flow = p->flows; flow; flow = flow->next)
if (flow->classid == classid) break;
return flow;
}
static int atm_tc_graft(struct Qdisc *sch,unsigned long arg,
struct Qdisc *new,struct Qdisc **old)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow = (struct atm_flow_data *) arg;
DPRINTK("atm_tc_graft(sch %p,[qdisc %p],flow %p,new %p,old %p)\n",sch,
p,flow,new,old);
if (!find_flow(p,flow)) return -EINVAL;
if (!new) new = &noop_qdisc;
*old = xchg(&flow->q,new);
if (*old) qdisc_reset(*old);
return 0;
}
static struct Qdisc *atm_tc_leaf(struct Qdisc *sch,unsigned long cl)
{
struct atm_flow_data *flow = (struct atm_flow_data *) cl;
DPRINTK("atm_tc_leaf(sch %p,flow %p)\n",sch,flow);
return flow ? flow->q : NULL;
}
static unsigned long atm_tc_get(struct Qdisc *sch,u32 classid)
{
struct atm_qdisc_data *p __attribute__((unused)) = PRIV(sch);
struct atm_flow_data *flow;
DPRINTK("atm_tc_get(sch %p,[qdisc %p],classid %x)\n",sch,p,classid);
flow = lookup_flow(sch,classid);
if (flow) flow->ref++;
DPRINTK("atm_tc_get: flow %p\n",flow);
return (unsigned long) flow;
}
static unsigned long atm_tc_bind_filter(struct Qdisc *sch,
unsigned long parent, u32 classid)
{
return atm_tc_get(sch,classid);
}
static void destroy_filters(struct atm_flow_data *flow)
{
struct tcf_proto *filter;
while ((filter = flow->filter_list)) {
DPRINTK("destroy_filters: destroying filter %p\n",filter);
flow->filter_list = filter->next;
tcf_destroy(filter);
}
}
/*
* atm_tc_put handles all destructions, including the ones that are explicitly
* requested (atm_tc_destroy, etc.). The assumption here is that we never drop
* anything that still seems to be in use.
*/
static void atm_tc_put(struct Qdisc *sch, unsigned long cl)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow = (struct atm_flow_data *) cl;
struct atm_flow_data **prev;
DPRINTK("atm_tc_put(sch %p,[qdisc %p],flow %p)\n",sch,p,flow);
if (--flow->ref) return;
DPRINTK("atm_tc_put: destroying\n");
for (prev = &p->flows; *prev; prev = &(*prev)->next)
if (*prev == flow) break;
if (!*prev) {
printk(KERN_CRIT "atm_tc_put: class %p not found\n",flow);
return;
}
*prev = flow->next;
DPRINTK("atm_tc_put: qdisc %p\n",flow->q);
qdisc_destroy(flow->q);
destroy_filters(flow);
if (flow->sock) {
DPRINTK("atm_tc_put: f_count %d\n",
file_count(flow->sock->file));
flow->vcc->pop = flow->old_pop;
sockfd_put(flow->sock);
}
if (flow->excess) atm_tc_put(sch,(unsigned long) flow->excess);
if (flow != &p->link) kfree(flow);
/*
* If flow == &p->link, the qdisc no longer works at this point and
* needs to be removed. (By the caller of atm_tc_put.)
*/
}
static void sch_atm_pop(struct atm_vcc *vcc,struct sk_buff *skb)
{
struct atm_qdisc_data *p = VCC2FLOW(vcc)->parent;
D2PRINTK("sch_atm_pop(vcc %p,skb %p,[qdisc %p])\n",vcc,skb,p);
VCC2FLOW(vcc)->old_pop(vcc,skb);
tasklet_schedule(&p->task);
}
static const u8 llc_oui_ip[] = {
0xaa, /* DSAP: non-ISO */
0xaa, /* SSAP: non-ISO */
0x03, /* Ctrl: Unnumbered Information Command PDU */
0x00, /* OUI: EtherType */
0x00, 0x00,
0x08, 0x00 }; /* Ethertype IP (0800) */
static int atm_tc_change(struct Qdisc *sch, u32 classid, u32 parent,
struct rtattr **tca, unsigned long *arg)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow = (struct atm_flow_data *) *arg;
struct atm_flow_data *excess = NULL;
struct rtattr *opt = tca[TCA_OPTIONS-1];
struct rtattr *tb[TCA_ATM_MAX];
struct socket *sock;
int fd,error,hdr_len;
void *hdr;
DPRINTK("atm_tc_change(sch %p,[qdisc %p],classid %x,parent %x,"
"flow %p,opt %p)\n",sch,p,classid,parent,flow,opt);
/*
* The concept of parents doesn't apply for this qdisc.
*/
if (parent && parent != TC_H_ROOT && parent != sch->handle)
return -EINVAL;
/*
* ATM classes cannot be changed. In order to change properties of the
* ATM connection, that socket needs to be modified directly (via the
* native ATM API. In order to send a flow to a different VC, the old
* class needs to be removed and a new one added. (This may be changed
* later.)
*/
if (flow) return -EBUSY;
if (opt == NULL || rtattr_parse_nested(tb, TCA_ATM_MAX, opt))
return -EINVAL;
if (!tb[TCA_ATM_FD-1] || RTA_PAYLOAD(tb[TCA_ATM_FD-1]) < sizeof(fd))
return -EINVAL;
fd = *(int *) RTA_DATA(tb[TCA_ATM_FD-1]);
DPRINTK("atm_tc_change: fd %d\n",fd);
if (tb[TCA_ATM_HDR-1]) {
hdr_len = RTA_PAYLOAD(tb[TCA_ATM_HDR-1]);
hdr = RTA_DATA(tb[TCA_ATM_HDR-1]);
}
else {
hdr_len = RFC1483LLC_LEN;
hdr = NULL; /* default LLC/SNAP for IP */
}
if (!tb[TCA_ATM_EXCESS-1]) excess = NULL;
else {
if (RTA_PAYLOAD(tb[TCA_ATM_EXCESS-1]) != sizeof(u32))
return -EINVAL;
excess = (struct atm_flow_data *) atm_tc_get(sch,
*(u32 *) RTA_DATA(tb[TCA_ATM_EXCESS-1]));
if (!excess) return -ENOENT;
}
DPRINTK("atm_tc_change: type %d, payload %d, hdr_len %d\n",
opt->rta_type,RTA_PAYLOAD(opt),hdr_len);
if (!(sock = sockfd_lookup(fd,&error))) return error; /* f_count++ */
DPRINTK("atm_tc_change: f_count %d\n",file_count(sock->file));
if (sock->ops->family != PF_ATMSVC && sock->ops->family != PF_ATMPVC) {
error = -EPROTOTYPE;
goto err_out;
}
/* @@@ should check if the socket is really operational or we'll crash
on vcc->send */
if (classid) {
if (TC_H_MAJ(classid ^ sch->handle)) {
DPRINTK("atm_tc_change: classid mismatch\n");
error = -EINVAL;
goto err_out;
}
if (find_flow(p,flow)) {
error = -EEXIST;
goto err_out;
}
}
else {
int i;
unsigned long cl;
for (i = 1; i < 0x8000; i++) {
classid = TC_H_MAKE(sch->handle,0x8000 | i);
if (!(cl = atm_tc_get(sch,classid))) break;
atm_tc_put(sch,cl);
}
}
DPRINTK("atm_tc_change: new id %x\n",classid);
flow = kmalloc(sizeof(struct atm_flow_data)+hdr_len,GFP_KERNEL);
DPRINTK("atm_tc_change: flow %p\n",flow);
if (!flow) {
error = -ENOBUFS;
goto err_out;
}
memset(flow,0,sizeof(*flow));
flow->filter_list = NULL;
if (!(flow->q = qdisc_create_dflt(sch->dev,&pfifo_qdisc_ops)))
flow->q = &noop_qdisc;
DPRINTK("atm_tc_change: qdisc %p\n",flow->q);
flow->sock = sock;
flow->vcc = ATM_SD(sock); /* speedup */
flow->vcc->user_back = flow;
DPRINTK("atm_tc_change: vcc %p\n",flow->vcc);
flow->old_pop = flow->vcc->pop;
flow->parent = p;
flow->vcc->pop = sch_atm_pop;
flow->classid = classid;
flow->ref = 1;
flow->excess = excess;
flow->next = p->link.next;
p->link.next = flow;
flow->hdr_len = hdr_len;
if (hdr)
memcpy(flow->hdr,hdr,hdr_len);
else
memcpy(flow->hdr,llc_oui_ip,sizeof(llc_oui_ip));
*arg = (unsigned long) flow;
return 0;
err_out:
if (excess) atm_tc_put(sch,(unsigned long) excess);
sockfd_put(sock);
return error;
}
static int atm_tc_delete(struct Qdisc *sch,unsigned long arg)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow = (struct atm_flow_data *) arg;
DPRINTK("atm_tc_delete(sch %p,[qdisc %p],flow %p)\n",sch,p,flow);
if (!find_flow(PRIV(sch),flow)) return -EINVAL;
if (flow->filter_list || flow == &p->link) return -EBUSY;
/*
* Reference count must be 2: one for "keepalive" (set at class
* creation), and one for the reference held when calling delete.
*/
if (flow->ref < 2) {
printk(KERN_ERR "atm_tc_delete: flow->ref == %d\n",flow->ref);
return -EINVAL;
}
if (flow->ref > 2) return -EBUSY; /* catch references via excess, etc.*/
atm_tc_put(sch,arg);
return 0;
}
static void atm_tc_walk(struct Qdisc *sch,struct qdisc_walker *walker)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow;
DPRINTK("atm_tc_walk(sch %p,[qdisc %p],walker %p)\n",sch,p,walker);
if (walker->stop) return;
for (flow = p->flows; flow; flow = flow->next) {
if (walker->count >= walker->skip)
if (walker->fn(sch,(unsigned long) flow,walker) < 0) {
walker->stop = 1;
break;
}
walker->count++;
}
}
static struct tcf_proto **atm_tc_find_tcf(struct Qdisc *sch,unsigned long cl)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow = (struct atm_flow_data *) cl;
DPRINTK("atm_tc_find_tcf(sch %p,[qdisc %p],flow %p)\n",sch,p,flow);
return flow ? &flow->filter_list : &p->link.filter_list;
}
/* --------------------------- Qdisc operations ---------------------------- */
static int atm_tc_enqueue(struct sk_buff *skb,struct Qdisc *sch)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow = NULL ; /* @@@ */
struct tcf_result res;
int result;
int ret = NET_XMIT_POLICED;
D2PRINTK("atm_tc_enqueue(skb %p,sch %p,[qdisc %p])\n",skb,sch,p);
result = TC_POLICE_OK; /* be nice to gcc */
if (TC_H_MAJ(skb->priority) != sch->handle ||
!(flow = (struct atm_flow_data *) atm_tc_get(sch,skb->priority)))
for (flow = p->flows; flow; flow = flow->next)
if (flow->filter_list) {
result = tc_classify(skb,flow->filter_list,
&res);
if (result < 0) continue;
flow = (struct atm_flow_data *) res.class;
if (!flow) flow = lookup_flow(sch,res.classid);
break;
}
if (!flow) flow = &p->link;
else {
if (flow->vcc)
ATM_SKB(skb)->atm_options = flow->vcc->atm_options;
/*@@@ looks good ... but it's not supposed to work :-)*/
#ifdef CONFIG_NET_CLS_POLICE
switch (result) {
case TC_POLICE_SHOT:
kfree_skb(skb);
break;
case TC_POLICE_RECLASSIFY:
if (flow->excess) flow = flow->excess;
else {
ATM_SKB(skb)->atm_options |=
ATM_ATMOPT_CLP;
break;
}
/* fall through */
case TC_POLICE_OK:
/* fall through */
default:
break;
}
#endif
}
if (
#ifdef CONFIG_NET_CLS_POLICE
result == TC_POLICE_SHOT ||
#endif
(ret = flow->q->enqueue(skb,flow->q)) != 0) {
sch->qstats.drops++;
if (flow) flow->qstats.drops++;
return ret;
}
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
flow->bstats.bytes += skb->len;
flow->bstats.packets++;
/*
* Okay, this may seem weird. We pretend we've dropped the packet if
* it goes via ATM. The reason for this is that the outer qdisc
* expects to be able to q->dequeue the packet later on if we return
* success at this place. Also, sch->q.qdisc needs to reflect whether
* there is a packet egligible for dequeuing or not. Note that the
* statistics of the outer qdisc are necessarily wrong because of all
* this. There's currently no correct solution for this.
*/
if (flow == &p->link) {
sch->q.qlen++;
return 0;
}
tasklet_schedule(&p->task);
return NET_XMIT_BYPASS;
}
/*
* Dequeue packets and send them over ATM. Note that we quite deliberately
* avoid checking net_device's flow control here, simply because sch_atm
* uses its own channels, which have nothing to do with any CLIP/LANE/or
* non-ATM interfaces.
*/
static void sch_atm_dequeue(unsigned long data)
{
struct Qdisc *sch = (struct Qdisc *) data;
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow;
struct sk_buff *skb;
D2PRINTK("sch_atm_dequeue(sch %p,[qdisc %p])\n",sch,p);
for (flow = p->link.next; flow; flow = flow->next)
/*
* If traffic is properly shaped, this won't generate nasty
* little bursts. Otherwise, it may ... (but that's okay)
*/
while ((skb = flow->q->dequeue(flow->q))) {
if (!atm_may_send(flow->vcc,skb->truesize)) {
(void) flow->q->ops->requeue(skb,flow->q);
break;
}
D2PRINTK("atm_tc_dequeue: sending on class %p\n",flow);
/* remove any LL header somebody else has attached */
skb_pull(skb,(char *) skb->nh.iph-(char *) skb->data);
if (skb_headroom(skb) < flow->hdr_len) {
struct sk_buff *new;
new = skb_realloc_headroom(skb,flow->hdr_len);
dev_kfree_skb(skb);
if (!new) continue;
skb = new;
}
D2PRINTK("sch_atm_dequeue: ip %p, data %p\n",
skb->nh.iph,skb->data);
ATM_SKB(skb)->vcc = flow->vcc;
memcpy(skb_push(skb,flow->hdr_len),flow->hdr,
flow->hdr_len);
atomic_add(skb->truesize,
&sk_atm(flow->vcc)->sk_wmem_alloc);
/* atm.atm_options are already set by atm_tc_enqueue */
(void) flow->vcc->send(flow->vcc,skb);
}
}
static struct sk_buff *atm_tc_dequeue(struct Qdisc *sch)
{
struct atm_qdisc_data *p = PRIV(sch);
struct sk_buff *skb;
D2PRINTK("atm_tc_dequeue(sch %p,[qdisc %p])\n",sch,p);
tasklet_schedule(&p->task);
skb = p->link.q->dequeue(p->link.q);
if (skb) sch->q.qlen--;
return skb;
}
static int atm_tc_requeue(struct sk_buff *skb,struct Qdisc *sch)
{
struct atm_qdisc_data *p = PRIV(sch);
int ret;
D2PRINTK("atm_tc_requeue(skb %p,sch %p,[qdisc %p])\n",skb,sch,p);
ret = p->link.q->ops->requeue(skb,p->link.q);
if (!ret) {
sch->q.qlen++;
sch->qstats.requeues++;
} else {
sch->qstats.drops++;
p->link.qstats.drops++;
}
return ret;
}
static unsigned int atm_tc_drop(struct Qdisc *sch)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow;
unsigned int len;
DPRINTK("atm_tc_drop(sch %p,[qdisc %p])\n",sch,p);
for (flow = p->flows; flow; flow = flow->next)
if (flow->q->ops->drop && (len = flow->q->ops->drop(flow->q)))
return len;
return 0;
}
static int atm_tc_init(struct Qdisc *sch,struct rtattr *opt)
{
struct atm_qdisc_data *p = PRIV(sch);
DPRINTK("atm_tc_init(sch %p,[qdisc %p],opt %p)\n",sch,p,opt);
p->flows = &p->link;
if(!(p->link.q = qdisc_create_dflt(sch->dev,&pfifo_qdisc_ops)))
p->link.q = &noop_qdisc;
DPRINTK("atm_tc_init: link (%p) qdisc %p\n",&p->link,p->link.q);
p->link.filter_list = NULL;
p->link.vcc = NULL;
p->link.sock = NULL;
p->link.classid = sch->handle;
p->link.ref = 1;
p->link.next = NULL;
tasklet_init(&p->task,sch_atm_dequeue,(unsigned long) sch);
return 0;
}
static void atm_tc_reset(struct Qdisc *sch)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow;
DPRINTK("atm_tc_reset(sch %p,[qdisc %p])\n",sch,p);
for (flow = p->flows; flow; flow = flow->next) qdisc_reset(flow->q);
sch->q.qlen = 0;
}
static void atm_tc_destroy(struct Qdisc *sch)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow;
DPRINTK("atm_tc_destroy(sch %p,[qdisc %p])\n",sch,p);
/* races ? */
while ((flow = p->flows)) {
destroy_filters(flow);
if (flow->ref > 1)
printk(KERN_ERR "atm_destroy: %p->ref = %d\n",flow,
flow->ref);
atm_tc_put(sch,(unsigned long) flow);
if (p->flows == flow) {
printk(KERN_ERR "atm_destroy: putting flow %p didn't "
"kill it\n",flow);
p->flows = flow->next; /* brute force */
break;
}
}
tasklet_kill(&p->task);
}
static int atm_tc_dump_class(struct Qdisc *sch, unsigned long cl,
struct sk_buff *skb, struct tcmsg *tcm)
{
struct atm_qdisc_data *p = PRIV(sch);
struct atm_flow_data *flow = (struct atm_flow_data *) cl;
unsigned char *b = skb->tail;
struct rtattr *rta;
DPRINTK("atm_tc_dump_class(sch %p,[qdisc %p],flow %p,skb %p,tcm %p)\n",
sch,p,flow,skb,tcm);
if (!find_flow(p,flow)) return -EINVAL;
tcm->tcm_handle = flow->classid;
rta = (struct rtattr *) b;
RTA_PUT(skb,TCA_OPTIONS,0,NULL);
RTA_PUT(skb,TCA_ATM_HDR,flow->hdr_len,flow->hdr);
if (flow->vcc) {
struct sockaddr_atmpvc pvc;
int state;
pvc.sap_family = AF_ATMPVC;
pvc.sap_addr.itf = flow->vcc->dev ? flow->vcc->dev->number : -1;
pvc.sap_addr.vpi = flow->vcc->vpi;
pvc.sap_addr.vci = flow->vcc->vci;
RTA_PUT(skb,TCA_ATM_ADDR,sizeof(pvc),&pvc);
state = ATM_VF2VS(flow->vcc->flags);
RTA_PUT(skb,TCA_ATM_STATE,sizeof(state),&state);
}
if (flow->excess)
RTA_PUT(skb,TCA_ATM_EXCESS,sizeof(u32),&flow->classid);
else {
static u32 zero;
RTA_PUT(skb,TCA_ATM_EXCESS,sizeof(zero),&zero);
}
rta->rta_len = skb->tail-b;
return skb->len;
rtattr_failure:
skb_trim(skb,b-skb->data);
return -1;
}
static int
atm_tc_dump_class_stats(struct Qdisc *sch, unsigned long arg,
struct gnet_dump *d)
{
struct atm_flow_data *flow = (struct atm_flow_data *) arg;
flow->qstats.qlen = flow->q->q.qlen;
if (gnet_stats_copy_basic(d, &flow->bstats) < 0 ||
gnet_stats_copy_queue(d, &flow->qstats) < 0)
return -1;
return 0;
}
static int atm_tc_dump(struct Qdisc *sch, struct sk_buff *skb)
{
return 0;
}
static struct Qdisc_class_ops atm_class_ops = {
.graft = atm_tc_graft,
.leaf = atm_tc_leaf,
.get = atm_tc_get,
.put = atm_tc_put,
.change = atm_tc_change,
.delete = atm_tc_delete,
.walk = atm_tc_walk,
.tcf_chain = atm_tc_find_tcf,
.bind_tcf = atm_tc_bind_filter,
.unbind_tcf = atm_tc_put,
.dump = atm_tc_dump_class,
.dump_stats = atm_tc_dump_class_stats,
};
static struct Qdisc_ops atm_qdisc_ops = {
.next = NULL,
.cl_ops = &atm_class_ops,
.id = "atm",
.priv_size = sizeof(struct atm_qdisc_data),
.enqueue = atm_tc_enqueue,
.dequeue = atm_tc_dequeue,
.requeue = atm_tc_requeue,
.drop = atm_tc_drop,
.init = atm_tc_init,
.reset = atm_tc_reset,
.destroy = atm_tc_destroy,
.change = NULL,
.dump = atm_tc_dump,
.owner = THIS_MODULE,
};
static int __init atm_init(void)
{
return register_qdisc(&atm_qdisc_ops);
}
static void __exit atm_exit(void)
{
unregister_qdisc(&atm_qdisc_ops);
}
module_init(atm_init)
module_exit(atm_exit)
MODULE_LICENSE("GPL");

2124
net/sched/sch_cbq.c Normal file
View File

File diff suppressed because it is too large Load Diff

479
net/sched/sch_dsmark.c Normal file
View File

@@ -0,0 +1,479 @@
/* net/sched/sch_dsmark.c - Differentiated Services field marker */
/* Written 1998-2000 by Werner Almesberger, EPFL ICA */
#include <linux/config.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/types.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/skbuff.h>
#include <linux/netdevice.h> /* for pkt_sched */
#include <linux/rtnetlink.h>
#include <net/pkt_sched.h>
#include <net/dsfield.h>
#include <net/inet_ecn.h>
#include <asm/byteorder.h>
#if 1 /* control */
#define DPRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define DPRINTK(format,args...)
#endif
#if 0 /* data */
#define D2PRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define D2PRINTK(format,args...)
#endif
#define PRIV(sch) qdisc_priv(sch)
/*
* classid class marking
* ------- ----- -------
* n/a 0 n/a
* x:0 1 use entry [0]
* ... ... ...
* x:y y>0 y+1 use entry [y]
* ... ... ...
* x:indices-1 indices use entry [indices-1]
* ... ... ...
* x:y y+1 use entry [y & (indices-1)]
* ... ... ...
* 0xffff 0x10000 use entry [indices-1]
*/
#define NO_DEFAULT_INDEX (1 << 16)
struct dsmark_qdisc_data {
struct Qdisc *q;
struct tcf_proto *filter_list;
__u8 *mask; /* "owns" the array */
__u8 *value;
__u16 indices;
__u32 default_index; /* index range is 0...0xffff */
int set_tc_index;
};
/* ------------------------- Class/flow operations ------------------------- */
static int dsmark_graft(struct Qdisc *sch,unsigned long arg,
struct Qdisc *new,struct Qdisc **old)
{
struct dsmark_qdisc_data *p = PRIV(sch);
DPRINTK("dsmark_graft(sch %p,[qdisc %p],new %p,old %p)\n",sch,p,new,
old);
if (!new)
new = &noop_qdisc;
sch_tree_lock(sch);
*old = xchg(&p->q,new);
if (*old)
qdisc_reset(*old);
sch->q.qlen = 0;
sch_tree_unlock(sch); /* @@@ move up ? */
return 0;
}
static struct Qdisc *dsmark_leaf(struct Qdisc *sch, unsigned long arg)
{
struct dsmark_qdisc_data *p = PRIV(sch);
return p->q;
}
static unsigned long dsmark_get(struct Qdisc *sch,u32 classid)
{
struct dsmark_qdisc_data *p __attribute__((unused)) = PRIV(sch);
DPRINTK("dsmark_get(sch %p,[qdisc %p],classid %x)\n",sch,p,classid);
return TC_H_MIN(classid)+1;
}
static unsigned long dsmark_bind_filter(struct Qdisc *sch,
unsigned long parent, u32 classid)
{
return dsmark_get(sch,classid);
}
static void dsmark_put(struct Qdisc *sch, unsigned long cl)
{
}
static int dsmark_change(struct Qdisc *sch, u32 classid, u32 parent,
struct rtattr **tca, unsigned long *arg)
{
struct dsmark_qdisc_data *p = PRIV(sch);
struct rtattr *opt = tca[TCA_OPTIONS-1];
struct rtattr *tb[TCA_DSMARK_MAX];
DPRINTK("dsmark_change(sch %p,[qdisc %p],classid %x,parent %x),"
"arg 0x%lx\n",sch,p,classid,parent,*arg);
if (*arg > p->indices)
return -ENOENT;
if (!opt || rtattr_parse_nested(tb, TCA_DSMARK_MAX, opt))
return -EINVAL;
if (tb[TCA_DSMARK_MASK-1]) {
if (!RTA_PAYLOAD(tb[TCA_DSMARK_MASK-1]))
return -EINVAL;
p->mask[*arg-1] = *(__u8 *) RTA_DATA(tb[TCA_DSMARK_MASK-1]);
}
if (tb[TCA_DSMARK_VALUE-1]) {
if (!RTA_PAYLOAD(tb[TCA_DSMARK_VALUE-1]))
return -EINVAL;
p->value[*arg-1] = *(__u8 *) RTA_DATA(tb[TCA_DSMARK_VALUE-1]);
}
return 0;
}
static int dsmark_delete(struct Qdisc *sch,unsigned long arg)
{
struct dsmark_qdisc_data *p = PRIV(sch);
if (!arg || arg > p->indices)
return -EINVAL;
p->mask[arg-1] = 0xff;
p->value[arg-1] = 0;
return 0;
}
static void dsmark_walk(struct Qdisc *sch,struct qdisc_walker *walker)
{
struct dsmark_qdisc_data *p = PRIV(sch);
int i;
DPRINTK("dsmark_walk(sch %p,[qdisc %p],walker %p)\n",sch,p,walker);
if (walker->stop)
return;
for (i = 0; i < p->indices; i++) {
if (p->mask[i] == 0xff && !p->value[i])
continue;
if (walker->count >= walker->skip) {
if (walker->fn(sch, i+1, walker) < 0) {
walker->stop = 1;
break;
}
}
walker->count++;
}
}
static struct tcf_proto **dsmark_find_tcf(struct Qdisc *sch,unsigned long cl)
{
struct dsmark_qdisc_data *p = PRIV(sch);
return &p->filter_list;
}
/* --------------------------- Qdisc operations ---------------------------- */
static int dsmark_enqueue(struct sk_buff *skb,struct Qdisc *sch)
{
struct dsmark_qdisc_data *p = PRIV(sch);
struct tcf_result res;
int result;
int ret = NET_XMIT_POLICED;
D2PRINTK("dsmark_enqueue(skb %p,sch %p,[qdisc %p])\n",skb,sch,p);
if (p->set_tc_index) {
/* FIXME: Safe with non-linear skbs? --RR */
switch (skb->protocol) {
case __constant_htons(ETH_P_IP):
skb->tc_index = ipv4_get_dsfield(skb->nh.iph)
& ~INET_ECN_MASK;
break;
case __constant_htons(ETH_P_IPV6):
skb->tc_index = ipv6_get_dsfield(skb->nh.ipv6h)
& ~INET_ECN_MASK;
break;
default:
skb->tc_index = 0;
break;
};
}
result = TC_POLICE_OK; /* be nice to gcc */
if (TC_H_MAJ(skb->priority) == sch->handle) {
skb->tc_index = TC_H_MIN(skb->priority);
} else {
result = tc_classify(skb,p->filter_list,&res);
D2PRINTK("result %d class 0x%04x\n",result,res.classid);
switch (result) {
#ifdef CONFIG_NET_CLS_POLICE
case TC_POLICE_SHOT:
kfree_skb(skb);
break;
#if 0
case TC_POLICE_RECLASSIFY:
/* FIXME: what to do here ??? */
#endif
#endif
case TC_POLICE_OK:
skb->tc_index = TC_H_MIN(res.classid);
break;
case TC_POLICE_UNSPEC:
/* fall through */
default:
if (p->default_index != NO_DEFAULT_INDEX)
skb->tc_index = p->default_index;
break;
};
}
if (
#ifdef CONFIG_NET_CLS_POLICE
result == TC_POLICE_SHOT ||
#endif
((ret = p->q->enqueue(skb,p->q)) != 0)) {
sch->qstats.drops++;
return ret;
}
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
sch->q.qlen++;
return ret;
}
static struct sk_buff *dsmark_dequeue(struct Qdisc *sch)
{
struct dsmark_qdisc_data *p = PRIV(sch);
struct sk_buff *skb;
int index;
D2PRINTK("dsmark_dequeue(sch %p,[qdisc %p])\n",sch,p);
skb = p->q->ops->dequeue(p->q);
if (!skb)
return NULL;
sch->q.qlen--;
index = skb->tc_index & (p->indices-1);
D2PRINTK("index %d->%d\n",skb->tc_index,index);
switch (skb->protocol) {
case __constant_htons(ETH_P_IP):
ipv4_change_dsfield(skb->nh.iph,
p->mask[index],p->value[index]);
break;
case __constant_htons(ETH_P_IPV6):
ipv6_change_dsfield(skb->nh.ipv6h,
p->mask[index],p->value[index]);
break;
default:
/*
* Only complain if a change was actually attempted.
* This way, we can send non-IP traffic through dsmark
* and don't need yet another qdisc as a bypass.
*/
if (p->mask[index] != 0xff || p->value[index])
printk(KERN_WARNING "dsmark_dequeue: "
"unsupported protocol %d\n",
htons(skb->protocol));
break;
};
return skb;
}
static int dsmark_requeue(struct sk_buff *skb,struct Qdisc *sch)
{
int ret;
struct dsmark_qdisc_data *p = PRIV(sch);
D2PRINTK("dsmark_requeue(skb %p,sch %p,[qdisc %p])\n",skb,sch,p);
if ((ret = p->q->ops->requeue(skb, p->q)) == 0) {
sch->q.qlen++;
sch->qstats.requeues++;
return 0;
}
sch->qstats.drops++;
return ret;
}
static unsigned int dsmark_drop(struct Qdisc *sch)
{
struct dsmark_qdisc_data *p = PRIV(sch);
unsigned int len;
DPRINTK("dsmark_reset(sch %p,[qdisc %p])\n",sch,p);
if (!p->q->ops->drop)
return 0;
if (!(len = p->q->ops->drop(p->q)))
return 0;
sch->q.qlen--;
return len;
}
static int dsmark_init(struct Qdisc *sch,struct rtattr *opt)
{
struct dsmark_qdisc_data *p = PRIV(sch);
struct rtattr *tb[TCA_DSMARK_MAX];
__u16 tmp;
DPRINTK("dsmark_init(sch %p,[qdisc %p],opt %p)\n",sch,p,opt);
if (!opt ||
rtattr_parse(tb,TCA_DSMARK_MAX,RTA_DATA(opt),RTA_PAYLOAD(opt)) < 0 ||
!tb[TCA_DSMARK_INDICES-1] ||
RTA_PAYLOAD(tb[TCA_DSMARK_INDICES-1]) < sizeof(__u16))
return -EINVAL;
p->indices = *(__u16 *) RTA_DATA(tb[TCA_DSMARK_INDICES-1]);
if (!p->indices)
return -EINVAL;
for (tmp = p->indices; tmp != 1; tmp >>= 1) {
if (tmp & 1)
return -EINVAL;
}
p->default_index = NO_DEFAULT_INDEX;
if (tb[TCA_DSMARK_DEFAULT_INDEX-1]) {
if (RTA_PAYLOAD(tb[TCA_DSMARK_DEFAULT_INDEX-1]) < sizeof(__u16))
return -EINVAL;
p->default_index =
*(__u16 *) RTA_DATA(tb[TCA_DSMARK_DEFAULT_INDEX-1]);
}
p->set_tc_index = !!tb[TCA_DSMARK_SET_TC_INDEX-1];
p->mask = kmalloc(p->indices*2,GFP_KERNEL);
if (!p->mask)
return -ENOMEM;
p->value = p->mask+p->indices;
memset(p->mask,0xff,p->indices);
memset(p->value,0,p->indices);
if (!(p->q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops)))
p->q = &noop_qdisc;
DPRINTK("dsmark_init: qdisc %p\n",&p->q);
return 0;
}
static void dsmark_reset(struct Qdisc *sch)
{
struct dsmark_qdisc_data *p = PRIV(sch);
DPRINTK("dsmark_reset(sch %p,[qdisc %p])\n",sch,p);
qdisc_reset(p->q);
sch->q.qlen = 0;
}
static void dsmark_destroy(struct Qdisc *sch)
{
struct dsmark_qdisc_data *p = PRIV(sch);
struct tcf_proto *tp;
DPRINTK("dsmark_destroy(sch %p,[qdisc %p])\n",sch,p);
while (p->filter_list) {
tp = p->filter_list;
p->filter_list = tp->next;
tcf_destroy(tp);
}
qdisc_destroy(p->q);
kfree(p->mask);
}
static int dsmark_dump_class(struct Qdisc *sch, unsigned long cl,
struct sk_buff *skb, struct tcmsg *tcm)
{
struct dsmark_qdisc_data *p = PRIV(sch);
unsigned char *b = skb->tail;
struct rtattr *rta;
DPRINTK("dsmark_dump_class(sch %p,[qdisc %p],class %ld\n",sch,p,cl);
if (!cl || cl > p->indices)
return -EINVAL;
tcm->tcm_handle = TC_H_MAKE(TC_H_MAJ(sch->handle),cl-1);
rta = (struct rtattr *) b;
RTA_PUT(skb,TCA_OPTIONS,0,NULL);
RTA_PUT(skb,TCA_DSMARK_MASK,1,&p->mask[cl-1]);
RTA_PUT(skb,TCA_DSMARK_VALUE,1,&p->value[cl-1]);
rta->rta_len = skb->tail-b;
return skb->len;
rtattr_failure:
skb_trim(skb,b-skb->data);
return -1;
}
static int dsmark_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct dsmark_qdisc_data *p = PRIV(sch);
unsigned char *b = skb->tail;
struct rtattr *rta;
rta = (struct rtattr *) b;
RTA_PUT(skb,TCA_OPTIONS,0,NULL);
RTA_PUT(skb,TCA_DSMARK_INDICES,sizeof(__u16),&p->indices);
if (p->default_index != NO_DEFAULT_INDEX) {
__u16 tmp = p->default_index;
RTA_PUT(skb,TCA_DSMARK_DEFAULT_INDEX, sizeof(__u16), &tmp);
}
if (p->set_tc_index)
RTA_PUT(skb, TCA_DSMARK_SET_TC_INDEX, 0, NULL);
rta->rta_len = skb->tail-b;
return skb->len;
rtattr_failure:
skb_trim(skb,b-skb->data);
return -1;
}
static struct Qdisc_class_ops dsmark_class_ops = {
.graft = dsmark_graft,
.leaf = dsmark_leaf,
.get = dsmark_get,
.put = dsmark_put,
.change = dsmark_change,
.delete = dsmark_delete,
.walk = dsmark_walk,
.tcf_chain = dsmark_find_tcf,
.bind_tcf = dsmark_bind_filter,
.unbind_tcf = dsmark_put,
.dump = dsmark_dump_class,
};
static struct Qdisc_ops dsmark_qdisc_ops = {
.next = NULL,
.cl_ops = &dsmark_class_ops,
.id = "dsmark",
.priv_size = sizeof(struct dsmark_qdisc_data),
.enqueue = dsmark_enqueue,
.dequeue = dsmark_dequeue,
.requeue = dsmark_requeue,
.drop = dsmark_drop,
.init = dsmark_init,
.reset = dsmark_reset,
.destroy = dsmark_destroy,
.change = NULL,
.dump = dsmark_dump,
.owner = THIS_MODULE,
};
static int __init dsmark_module_init(void)
{
return register_qdisc(&dsmark_qdisc_ops);
}
static void __exit dsmark_module_exit(void)
{
unregister_qdisc(&dsmark_qdisc_ops);
}
module_init(dsmark_module_init)
module_exit(dsmark_module_exit)
MODULE_LICENSE("GPL");

212
net/sched/sch_fifo.c Normal file
View File

@@ -0,0 +1,212 @@
/*
* net/sched/sch_fifo.c The simplest FIFO queue.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
#include <linux/config.h>
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
/* 1 band FIFO pseudo-"scheduler" */
struct fifo_sched_data
{
unsigned limit;
};
static int
bfifo_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct fifo_sched_data *q = qdisc_priv(sch);
if (sch->qstats.backlog + skb->len <= q->limit) {
__skb_queue_tail(&sch->q, skb);
sch->qstats.backlog += skb->len;
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
return 0;
}
sch->qstats.drops++;
#ifdef CONFIG_NET_CLS_POLICE
if (sch->reshape_fail==NULL || sch->reshape_fail(skb, sch))
#endif
kfree_skb(skb);
return NET_XMIT_DROP;
}
static int
bfifo_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
__skb_queue_head(&sch->q, skb);
sch->qstats.backlog += skb->len;
sch->qstats.requeues++;
return 0;
}
static struct sk_buff *
bfifo_dequeue(struct Qdisc* sch)
{
struct sk_buff *skb;
skb = __skb_dequeue(&sch->q);
if (skb)
sch->qstats.backlog -= skb->len;
return skb;
}
static unsigned int
fifo_drop(struct Qdisc* sch)
{
struct sk_buff *skb;
skb = __skb_dequeue_tail(&sch->q);
if (skb) {
unsigned int len = skb->len;
sch->qstats.backlog -= len;
kfree_skb(skb);
return len;
}
return 0;
}
static void
fifo_reset(struct Qdisc* sch)
{
skb_queue_purge(&sch->q);
sch->qstats.backlog = 0;
}
static int
pfifo_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct fifo_sched_data *q = qdisc_priv(sch);
if (sch->q.qlen < q->limit) {
__skb_queue_tail(&sch->q, skb);
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
return 0;
}
sch->qstats.drops++;
#ifdef CONFIG_NET_CLS_POLICE
if (sch->reshape_fail==NULL || sch->reshape_fail(skb, sch))
#endif
kfree_skb(skb);
return NET_XMIT_DROP;
}
static int
pfifo_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
__skb_queue_head(&sch->q, skb);
sch->qstats.requeues++;
return 0;
}
static struct sk_buff *
pfifo_dequeue(struct Qdisc* sch)
{
return __skb_dequeue(&sch->q);
}
static int fifo_init(struct Qdisc *sch, struct rtattr *opt)
{
struct fifo_sched_data *q = qdisc_priv(sch);
if (opt == NULL) {
unsigned int limit = sch->dev->tx_queue_len ? : 1;
if (sch->ops == &bfifo_qdisc_ops)
q->limit = limit*sch->dev->mtu;
else
q->limit = limit;
} else {
struct tc_fifo_qopt *ctl = RTA_DATA(opt);
if (opt->rta_len < RTA_LENGTH(sizeof(*ctl)))
return -EINVAL;
q->limit = ctl->limit;
}
return 0;
}
static int fifo_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct fifo_sched_data *q = qdisc_priv(sch);
unsigned char *b = skb->tail;
struct tc_fifo_qopt opt;
opt.limit = q->limit;
RTA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
struct Qdisc_ops pfifo_qdisc_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "pfifo",
.priv_size = sizeof(struct fifo_sched_data),
.enqueue = pfifo_enqueue,
.dequeue = pfifo_dequeue,
.requeue = pfifo_requeue,
.drop = fifo_drop,
.init = fifo_init,
.reset = fifo_reset,
.destroy = NULL,
.change = fifo_init,
.dump = fifo_dump,
.owner = THIS_MODULE,
};
struct Qdisc_ops bfifo_qdisc_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "bfifo",
.priv_size = sizeof(struct fifo_sched_data),
.enqueue = bfifo_enqueue,
.dequeue = bfifo_dequeue,
.requeue = bfifo_requeue,
.drop = fifo_drop,
.init = fifo_init,
.reset = fifo_reset,
.destroy = NULL,
.change = fifo_init,
.dump = fifo_dump,
.owner = THIS_MODULE,
};
EXPORT_SYMBOL(bfifo_qdisc_ops);
EXPORT_SYMBOL(pfifo_qdisc_ops);

609
net/sched/sch_generic.c Normal file
View File

@@ -0,0 +1,609 @@
/*
* net/sched/sch_generic.c Generic packet scheduler routines.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
* Jamal Hadi Salim, <hadi@cyberus.ca> 990601
* - Ingress support
*/
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/init.h>
#include <linux/rcupdate.h>
#include <linux/list.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
/* Main transmission queue. */
/* Main qdisc structure lock.
However, modifications
to data, participating in scheduling must be additionally
protected with dev->queue_lock spinlock.
The idea is the following:
- enqueue, dequeue are serialized via top level device
spinlock dev->queue_lock.
- tree walking is protected by read_lock_bh(qdisc_tree_lock)
and this lock is used only in process context.
- updates to tree are made under rtnl semaphore or
from softirq context (__qdisc_destroy rcu-callback)
hence this lock needs local bh disabling.
qdisc_tree_lock must be grabbed BEFORE dev->queue_lock!
*/
DEFINE_RWLOCK(qdisc_tree_lock);
void qdisc_lock_tree(struct net_device *dev)
{
write_lock_bh(&qdisc_tree_lock);
spin_lock_bh(&dev->queue_lock);
}
void qdisc_unlock_tree(struct net_device *dev)
{
spin_unlock_bh(&dev->queue_lock);
write_unlock_bh(&qdisc_tree_lock);
}
/*
dev->queue_lock serializes queue accesses for this device
AND dev->qdisc pointer itself.
dev->xmit_lock serializes accesses to device driver.
dev->queue_lock and dev->xmit_lock are mutually exclusive,
if one is grabbed, another must be free.
*/
/* Kick device.
Note, that this procedure can be called by a watchdog timer, so that
we do not check dev->tbusy flag here.
Returns: 0 - queue is empty.
>0 - queue is not empty, but throttled.
<0 - queue is not empty. Device is throttled, if dev->tbusy != 0.
NOTE: Called under dev->queue_lock with locally disabled BH.
*/
int qdisc_restart(struct net_device *dev)
{
struct Qdisc *q = dev->qdisc;
struct sk_buff *skb;
/* Dequeue packet */
if ((skb = q->dequeue(q)) != NULL) {
unsigned nolock = (dev->features & NETIF_F_LLTX);
/*
* When the driver has LLTX set it does its own locking
* in start_xmit. No need to add additional overhead by
* locking again. These checks are worth it because
* even uncongested locks can be quite expensive.
* The driver can do trylock like here too, in case
* of lock congestion it should return -1 and the packet
* will be requeued.
*/
if (!nolock) {
if (!spin_trylock(&dev->xmit_lock)) {
collision:
/* So, someone grabbed the driver. */
/* It may be transient configuration error,
when hard_start_xmit() recurses. We detect
it by checking xmit owner and drop the
packet when deadloop is detected.
*/
if (dev->xmit_lock_owner == smp_processor_id()) {
kfree_skb(skb);
if (net_ratelimit())
printk(KERN_DEBUG "Dead loop on netdevice %s, fix it urgently!\n", dev->name);
return -1;
}
__get_cpu_var(netdev_rx_stat).cpu_collision++;
goto requeue;
}
/* Remember that the driver is grabbed by us. */
dev->xmit_lock_owner = smp_processor_id();
}
{
/* And release queue */
spin_unlock(&dev->queue_lock);
if (!netif_queue_stopped(dev)) {
int ret;
if (netdev_nit)
dev_queue_xmit_nit(skb, dev);
ret = dev->hard_start_xmit(skb, dev);
if (ret == NETDEV_TX_OK) {
if (!nolock) {
dev->xmit_lock_owner = -1;
spin_unlock(&dev->xmit_lock);
}
spin_lock(&dev->queue_lock);
return -1;
}
if (ret == NETDEV_TX_LOCKED && nolock) {
spin_lock(&dev->queue_lock);
goto collision;
}
}
/* NETDEV_TX_BUSY - we need to requeue */
/* Release the driver */
if (!nolock) {
dev->xmit_lock_owner = -1;
spin_unlock(&dev->xmit_lock);
}
spin_lock(&dev->queue_lock);
q = dev->qdisc;
}
/* Device kicked us out :(
This is possible in three cases:
0. driver is locked
1. fastroute is enabled
2. device cannot determine busy state
before start of transmission (f.e. dialout)
3. device is buggy (ppp)
*/
requeue:
q->ops->requeue(skb, q);
netif_schedule(dev);
return 1;
}
return q->q.qlen;
}
static void dev_watchdog(unsigned long arg)
{
struct net_device *dev = (struct net_device *)arg;
spin_lock(&dev->xmit_lock);
if (dev->qdisc != &noop_qdisc) {
if (netif_device_present(dev) &&
netif_running(dev) &&
netif_carrier_ok(dev)) {
if (netif_queue_stopped(dev) &&
(jiffies - dev->trans_start) > dev->watchdog_timeo) {
printk(KERN_INFO "NETDEV WATCHDOG: %s: transmit timed out\n", dev->name);
dev->tx_timeout(dev);
}
if (!mod_timer(&dev->watchdog_timer, jiffies + dev->watchdog_timeo))
dev_hold(dev);
}
}
spin_unlock(&dev->xmit_lock);
dev_put(dev);
}
static void dev_watchdog_init(struct net_device *dev)
{
init_timer(&dev->watchdog_timer);
dev->watchdog_timer.data = (unsigned long)dev;
dev->watchdog_timer.function = dev_watchdog;
}
void __netdev_watchdog_up(struct net_device *dev)
{
if (dev->tx_timeout) {
if (dev->watchdog_timeo <= 0)
dev->watchdog_timeo = 5*HZ;
if (!mod_timer(&dev->watchdog_timer, jiffies + dev->watchdog_timeo))
dev_hold(dev);
}
}
static void dev_watchdog_up(struct net_device *dev)
{
spin_lock_bh(&dev->xmit_lock);
__netdev_watchdog_up(dev);
spin_unlock_bh(&dev->xmit_lock);
}
static void dev_watchdog_down(struct net_device *dev)
{
spin_lock_bh(&dev->xmit_lock);
if (del_timer(&dev->watchdog_timer))
__dev_put(dev);
spin_unlock_bh(&dev->xmit_lock);
}
/* "NOOP" scheduler: the best scheduler, recommended for all interfaces
under all circumstances. It is difficult to invent anything faster or
cheaper.
*/
static int
noop_enqueue(struct sk_buff *skb, struct Qdisc * qdisc)
{
kfree_skb(skb);
return NET_XMIT_CN;
}
static struct sk_buff *
noop_dequeue(struct Qdisc * qdisc)
{
return NULL;
}
static int
noop_requeue(struct sk_buff *skb, struct Qdisc* qdisc)
{
if (net_ratelimit())
printk(KERN_DEBUG "%s deferred output. It is buggy.\n", skb->dev->name);
kfree_skb(skb);
return NET_XMIT_CN;
}
struct Qdisc_ops noop_qdisc_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "noop",
.priv_size = 0,
.enqueue = noop_enqueue,
.dequeue = noop_dequeue,
.requeue = noop_requeue,
.owner = THIS_MODULE,
};
struct Qdisc noop_qdisc = {
.enqueue = noop_enqueue,
.dequeue = noop_dequeue,
.flags = TCQ_F_BUILTIN,
.ops = &noop_qdisc_ops,
.list = LIST_HEAD_INIT(noop_qdisc.list),
};
static struct Qdisc_ops noqueue_qdisc_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "noqueue",
.priv_size = 0,
.enqueue = noop_enqueue,
.dequeue = noop_dequeue,
.requeue = noop_requeue,
.owner = THIS_MODULE,
};
static struct Qdisc noqueue_qdisc = {
.enqueue = NULL,
.dequeue = noop_dequeue,
.flags = TCQ_F_BUILTIN,
.ops = &noqueue_qdisc_ops,
.list = LIST_HEAD_INIT(noqueue_qdisc.list),
};
static const u8 prio2band[TC_PRIO_MAX+1] =
{ 1, 2, 2, 2, 1, 2, 0, 0 , 1, 1, 1, 1, 1, 1, 1, 1 };
/* 3-band FIFO queue: old style, but should be a bit faster than
generic prio+fifo combination.
*/
static int
pfifo_fast_enqueue(struct sk_buff *skb, struct Qdisc* qdisc)
{
struct sk_buff_head *list = qdisc_priv(qdisc);
list += prio2band[skb->priority&TC_PRIO_MAX];
if (list->qlen < qdisc->dev->tx_queue_len) {
__skb_queue_tail(list, skb);
qdisc->q.qlen++;
qdisc->bstats.bytes += skb->len;
qdisc->bstats.packets++;
return 0;
}
qdisc->qstats.drops++;
kfree_skb(skb);
return NET_XMIT_DROP;
}
static struct sk_buff *
pfifo_fast_dequeue(struct Qdisc* qdisc)
{
int prio;
struct sk_buff_head *list = qdisc_priv(qdisc);
struct sk_buff *skb;
for (prio = 0; prio < 3; prio++, list++) {
skb = __skb_dequeue(list);
if (skb) {
qdisc->q.qlen--;
return skb;
}
}
return NULL;
}
static int
pfifo_fast_requeue(struct sk_buff *skb, struct Qdisc* qdisc)
{
struct sk_buff_head *list = qdisc_priv(qdisc);
list += prio2band[skb->priority&TC_PRIO_MAX];
__skb_queue_head(list, skb);
qdisc->q.qlen++;
qdisc->qstats.requeues++;
return 0;
}
static void
pfifo_fast_reset(struct Qdisc* qdisc)
{
int prio;
struct sk_buff_head *list = qdisc_priv(qdisc);
for (prio=0; prio < 3; prio++)
skb_queue_purge(list+prio);
qdisc->q.qlen = 0;
}
static int pfifo_fast_dump(struct Qdisc *qdisc, struct sk_buff *skb)
{
unsigned char *b = skb->tail;
struct tc_prio_qopt opt;
opt.bands = 3;
memcpy(&opt.priomap, prio2band, TC_PRIO_MAX+1);
RTA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int pfifo_fast_init(struct Qdisc *qdisc, struct rtattr *opt)
{
int i;
struct sk_buff_head *list = qdisc_priv(qdisc);
for (i=0; i<3; i++)
skb_queue_head_init(list+i);
return 0;
}
static struct Qdisc_ops pfifo_fast_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "pfifo_fast",
.priv_size = 3 * sizeof(struct sk_buff_head),
.enqueue = pfifo_fast_enqueue,
.dequeue = pfifo_fast_dequeue,
.requeue = pfifo_fast_requeue,
.init = pfifo_fast_init,
.reset = pfifo_fast_reset,
.dump = pfifo_fast_dump,
.owner = THIS_MODULE,
};
struct Qdisc * qdisc_create_dflt(struct net_device *dev, struct Qdisc_ops *ops)
{
void *p;
struct Qdisc *sch;
int size;
/* ensure that the Qdisc and the private data are 32-byte aligned */
size = ((sizeof(*sch) + QDISC_ALIGN_CONST) & ~QDISC_ALIGN_CONST);
size += ops->priv_size + QDISC_ALIGN_CONST;
p = kmalloc(size, GFP_KERNEL);
if (!p)
return NULL;
memset(p, 0, size);
sch = (struct Qdisc *)(((unsigned long)p + QDISC_ALIGN_CONST)
& ~QDISC_ALIGN_CONST);
sch->padded = (char *)sch - (char *)p;
INIT_LIST_HEAD(&sch->list);
skb_queue_head_init(&sch->q);
sch->ops = ops;
sch->enqueue = ops->enqueue;
sch->dequeue = ops->dequeue;
sch->dev = dev;
dev_hold(dev);
sch->stats_lock = &dev->queue_lock;
atomic_set(&sch->refcnt, 1);
if (!ops->init || ops->init(sch, NULL) == 0)
return sch;
dev_put(dev);
kfree(p);
return NULL;
}
/* Under dev->queue_lock and BH! */
void qdisc_reset(struct Qdisc *qdisc)
{
struct Qdisc_ops *ops = qdisc->ops;
if (ops->reset)
ops->reset(qdisc);
}
/* this is the rcu callback function to clean up a qdisc when there
* are no further references to it */
static void __qdisc_destroy(struct rcu_head *head)
{
struct Qdisc *qdisc = container_of(head, struct Qdisc, q_rcu);
struct Qdisc_ops *ops = qdisc->ops;
#ifdef CONFIG_NET_ESTIMATOR
gen_kill_estimator(&qdisc->bstats, &qdisc->rate_est);
#endif
write_lock(&qdisc_tree_lock);
if (ops->reset)
ops->reset(qdisc);
if (ops->destroy)
ops->destroy(qdisc);
write_unlock(&qdisc_tree_lock);
module_put(ops->owner);
dev_put(qdisc->dev);
kfree((char *) qdisc - qdisc->padded);
}
/* Under dev->queue_lock and BH! */
void qdisc_destroy(struct Qdisc *qdisc)
{
struct list_head cql = LIST_HEAD_INIT(cql);
struct Qdisc *cq, *q, *n;
if (qdisc->flags & TCQ_F_BUILTIN ||
!atomic_dec_and_test(&qdisc->refcnt))
return;
if (!list_empty(&qdisc->list)) {
if (qdisc->ops->cl_ops == NULL)
list_del(&qdisc->list);
else
list_move(&qdisc->list, &cql);
}
/* unlink inner qdiscs from dev->qdisc_list immediately */
list_for_each_entry(cq, &cql, list)
list_for_each_entry_safe(q, n, &qdisc->dev->qdisc_list, list)
if (TC_H_MAJ(q->parent) == TC_H_MAJ(cq->handle)) {
if (q->ops->cl_ops == NULL)
list_del_init(&q->list);
else
list_move_tail(&q->list, &cql);
}
list_for_each_entry_safe(cq, n, &cql, list)
list_del_init(&cq->list);
call_rcu(&qdisc->q_rcu, __qdisc_destroy);
}
void dev_activate(struct net_device *dev)
{
/* No queueing discipline is attached to device;
create default one i.e. pfifo_fast for devices,
which need queueing and noqueue_qdisc for
virtual interfaces
*/
if (dev->qdisc_sleeping == &noop_qdisc) {
struct Qdisc *qdisc;
if (dev->tx_queue_len) {
qdisc = qdisc_create_dflt(dev, &pfifo_fast_ops);
if (qdisc == NULL) {
printk(KERN_INFO "%s: activation failed\n", dev->name);
return;
}
write_lock_bh(&qdisc_tree_lock);
list_add_tail(&qdisc->list, &dev->qdisc_list);
write_unlock_bh(&qdisc_tree_lock);
} else {
qdisc = &noqueue_qdisc;
}
write_lock_bh(&qdisc_tree_lock);
dev->qdisc_sleeping = qdisc;
write_unlock_bh(&qdisc_tree_lock);
}
spin_lock_bh(&dev->queue_lock);
rcu_assign_pointer(dev->qdisc, dev->qdisc_sleeping);
if (dev->qdisc != &noqueue_qdisc) {
dev->trans_start = jiffies;
dev_watchdog_up(dev);
}
spin_unlock_bh(&dev->queue_lock);
}
void dev_deactivate(struct net_device *dev)
{
struct Qdisc *qdisc;
spin_lock_bh(&dev->queue_lock);
qdisc = dev->qdisc;
dev->qdisc = &noop_qdisc;
qdisc_reset(qdisc);
spin_unlock_bh(&dev->queue_lock);
dev_watchdog_down(dev);
while (test_bit(__LINK_STATE_SCHED, &dev->state))
yield();
spin_unlock_wait(&dev->xmit_lock);
}
void dev_init_scheduler(struct net_device *dev)
{
qdisc_lock_tree(dev);
dev->qdisc = &noop_qdisc;
dev->qdisc_sleeping = &noop_qdisc;
INIT_LIST_HEAD(&dev->qdisc_list);
qdisc_unlock_tree(dev);
dev_watchdog_init(dev);
}
void dev_shutdown(struct net_device *dev)
{
struct Qdisc *qdisc;
qdisc_lock_tree(dev);
qdisc = dev->qdisc_sleeping;
dev->qdisc = &noop_qdisc;
dev->qdisc_sleeping = &noop_qdisc;
qdisc_destroy(qdisc);
#if defined(CONFIG_NET_SCH_INGRESS) || defined(CONFIG_NET_SCH_INGRESS_MODULE)
if ((qdisc = dev->qdisc_ingress) != NULL) {
dev->qdisc_ingress = NULL;
qdisc_destroy(qdisc);
}
#endif
BUG_TRAP(!timer_pending(&dev->watchdog_timer));
qdisc_unlock_tree(dev);
}
EXPORT_SYMBOL(__netdev_watchdog_up);
EXPORT_SYMBOL(noop_qdisc);
EXPORT_SYMBOL(noop_qdisc_ops);
EXPORT_SYMBOL(qdisc_create_dflt);
EXPORT_SYMBOL(qdisc_destroy);
EXPORT_SYMBOL(qdisc_reset);
EXPORT_SYMBOL(qdisc_restart);
EXPORT_SYMBOL(qdisc_lock_tree);
EXPORT_SYMBOL(qdisc_unlock_tree);

630
net/sched/sch_gred.c Normal file
View File

@@ -0,0 +1,630 @@
/*
* net/sched/sch_gred.c Generic Random Early Detection queue.
*
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: J Hadi Salim (hadi@cyberus.ca) 1998-2002
*
* 991129: - Bug fix with grio mode
* - a better sing. AvgQ mode with Grio(WRED)
* - A finer grained VQ dequeue based on sugestion
* from Ren Liu
* - More error checks
*
*
*
* For all the glorious comments look at Alexey's sch_red.c
*/
#include <linux/config.h>
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
#if 1 /* control */
#define DPRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define DPRINTK(format,args...)
#endif
#if 0 /* data */
#define D2PRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define D2PRINTK(format,args...)
#endif
struct gred_sched_data;
struct gred_sched;
struct gred_sched_data
{
/* Parameters */
u32 limit; /* HARD maximal queue length */
u32 qth_min; /* Min average length threshold: A scaled */
u32 qth_max; /* Max average length threshold: A scaled */
u32 DP; /* the drop pramaters */
char Wlog; /* log(W) */
char Plog; /* random number bits */
u32 Scell_max;
u32 Rmask;
u32 bytesin; /* bytes seen on virtualQ so far*/
u32 packetsin; /* packets seen on virtualQ so far*/
u32 backlog; /* bytes on the virtualQ */
u32 forced; /* packets dropped for exceeding limits */
u32 early; /* packets dropped as a warning */
u32 other; /* packets dropped by invoking drop() */
u32 pdrop; /* packets dropped because we exceeded physical queue limits */
char Scell_log;
u8 Stab[256];
u8 prio; /* the prio of this vq */
/* Variables */
unsigned long qave; /* Average queue length: A scaled */
int qcount; /* Packets since last random number generation */
u32 qR; /* Cached random number */
psched_time_t qidlestart; /* Start of idle period */
};
struct gred_sched
{
struct gred_sched_data *tab[MAX_DPs];
u32 DPs;
u32 def;
u8 initd;
u8 grio;
u8 eqp;
};
static int
gred_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
psched_time_t now;
struct gred_sched_data *q=NULL;
struct gred_sched *t= qdisc_priv(sch);
unsigned long qave=0;
int i=0;
if (!t->initd && skb_queue_len(&sch->q) < (sch->dev->tx_queue_len ? : 1)) {
D2PRINTK("NO GRED Queues setup yet! Enqueued anyway\n");
goto do_enqueue;
}
if ( ((skb->tc_index&0xf) > (t->DPs -1)) || !(q=t->tab[skb->tc_index&0xf])) {
printk("GRED: setting to default (%d)\n ",t->def);
if (!(q=t->tab[t->def])) {
DPRINTK("GRED: setting to default FAILED! dropping!! "
"(%d)\n ", t->def);
goto drop;
}
/* fix tc_index? --could be controvesial but needed for
requeueing */
skb->tc_index=(skb->tc_index&0xfffffff0) | t->def;
}
D2PRINTK("gred_enqueue virtualQ 0x%x classid %x backlog %d "
"general backlog %d\n",skb->tc_index&0xf,sch->handle,q->backlog,
sch->qstats.backlog);
/* sum up all the qaves of prios <= to ours to get the new qave*/
if (!t->eqp && t->grio) {
for (i=0;i<t->DPs;i++) {
if ((!t->tab[i]) || (i==q->DP))
continue;
if ((t->tab[i]->prio < q->prio) && (PSCHED_IS_PASTPERFECT(t->tab[i]->qidlestart)))
qave +=t->tab[i]->qave;
}
}
q->packetsin++;
q->bytesin+=skb->len;
if (t->eqp && t->grio) {
qave=0;
q->qave=t->tab[t->def]->qave;
q->qidlestart=t->tab[t->def]->qidlestart;
}
if (!PSCHED_IS_PASTPERFECT(q->qidlestart)) {
long us_idle;
PSCHED_GET_TIME(now);
us_idle = PSCHED_TDIFF_SAFE(now, q->qidlestart, q->Scell_max);
PSCHED_SET_PASTPERFECT(q->qidlestart);
q->qave >>= q->Stab[(us_idle>>q->Scell_log)&0xFF];
} else {
if (t->eqp) {
q->qave += sch->qstats.backlog - (q->qave >> q->Wlog);
} else {
q->qave += q->backlog - (q->qave >> q->Wlog);
}
}
if (t->eqp && t->grio)
t->tab[t->def]->qave=q->qave;
if ((q->qave+qave) < q->qth_min) {
q->qcount = -1;
enqueue:
if (q->backlog + skb->len <= q->limit) {
q->backlog += skb->len;
do_enqueue:
__skb_queue_tail(&sch->q, skb);
sch->qstats.backlog += skb->len;
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
return 0;
} else {
q->pdrop++;
}
drop:
kfree_skb(skb);
sch->qstats.drops++;
return NET_XMIT_DROP;
}
if ((q->qave+qave) >= q->qth_max) {
q->qcount = -1;
sch->qstats.overlimits++;
q->forced++;
goto drop;
}
if (++q->qcount) {
if ((((qave+q->qave) - q->qth_min)>>q->Wlog)*q->qcount < q->qR)
goto enqueue;
q->qcount = 0;
q->qR = net_random()&q->Rmask;
sch->qstats.overlimits++;
q->early++;
goto drop;
}
q->qR = net_random()&q->Rmask;
goto enqueue;
}
static int
gred_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
struct gred_sched_data *q;
struct gred_sched *t= qdisc_priv(sch);
q= t->tab[(skb->tc_index&0xf)];
/* error checking here -- probably unnecessary */
PSCHED_SET_PASTPERFECT(q->qidlestart);
__skb_queue_head(&sch->q, skb);
sch->qstats.backlog += skb->len;
sch->qstats.requeues++;
q->backlog += skb->len;
return 0;
}
static struct sk_buff *
gred_dequeue(struct Qdisc* sch)
{
struct sk_buff *skb;
struct gred_sched_data *q;
struct gred_sched *t= qdisc_priv(sch);
skb = __skb_dequeue(&sch->q);
if (skb) {
sch->qstats.backlog -= skb->len;
q= t->tab[(skb->tc_index&0xf)];
if (q) {
q->backlog -= skb->len;
if (!q->backlog && !t->eqp)
PSCHED_GET_TIME(q->qidlestart);
} else {
D2PRINTK("gred_dequeue: skb has bad tcindex %x\n",skb->tc_index&0xf);
}
return skb;
}
if (t->eqp) {
q= t->tab[t->def];
if (!q)
D2PRINTK("no default VQ set: Results will be "
"screwed up\n");
else
PSCHED_GET_TIME(q->qidlestart);
}
return NULL;
}
static unsigned int gred_drop(struct Qdisc* sch)
{
struct sk_buff *skb;
struct gred_sched_data *q;
struct gred_sched *t= qdisc_priv(sch);
skb = __skb_dequeue_tail(&sch->q);
if (skb) {
unsigned int len = skb->len;
sch->qstats.backlog -= len;
sch->qstats.drops++;
q= t->tab[(skb->tc_index&0xf)];
if (q) {
q->backlog -= len;
q->other++;
if (!q->backlog && !t->eqp)
PSCHED_GET_TIME(q->qidlestart);
} else {
D2PRINTK("gred_dequeue: skb has bad tcindex %x\n",skb->tc_index&0xf);
}
kfree_skb(skb);
return len;
}
q=t->tab[t->def];
if (!q) {
D2PRINTK("no default VQ set: Results might be screwed up\n");
return 0;
}
PSCHED_GET_TIME(q->qidlestart);
return 0;
}
static void gred_reset(struct Qdisc* sch)
{
int i;
struct gred_sched_data *q;
struct gred_sched *t= qdisc_priv(sch);
__skb_queue_purge(&sch->q);
sch->qstats.backlog = 0;
for (i=0;i<t->DPs;i++) {
q= t->tab[i];
if (!q)
continue;
PSCHED_SET_PASTPERFECT(q->qidlestart);
q->qave = 0;
q->qcount = -1;
q->backlog = 0;
q->other=0;
q->forced=0;
q->pdrop=0;
q->early=0;
}
}
static int gred_change(struct Qdisc *sch, struct rtattr *opt)
{
struct gred_sched *table = qdisc_priv(sch);
struct gred_sched_data *q;
struct tc_gred_qopt *ctl;
struct tc_gred_sopt *sopt;
struct rtattr *tb[TCA_GRED_STAB];
struct rtattr *tb2[TCA_GRED_DPS];
int i;
if (opt == NULL || rtattr_parse_nested(tb, TCA_GRED_STAB, opt))
return -EINVAL;
if (tb[TCA_GRED_PARMS-1] == 0 && tb[TCA_GRED_STAB-1] == 0) {
rtattr_parse_nested(tb2, TCA_GRED_DPS, opt);
if (tb2[TCA_GRED_DPS-1] == 0)
return -EINVAL;
sopt = RTA_DATA(tb2[TCA_GRED_DPS-1]);
table->DPs=sopt->DPs;
table->def=sopt->def_DP;
table->grio=sopt->grio;
table->initd=0;
/* probably need to clear all the table DP entries as well */
return 0;
}
if (!table->DPs || tb[TCA_GRED_PARMS-1] == 0 || tb[TCA_GRED_STAB-1] == 0 ||
RTA_PAYLOAD(tb[TCA_GRED_PARMS-1]) < sizeof(*ctl) ||
RTA_PAYLOAD(tb[TCA_GRED_STAB-1]) < 256)
return -EINVAL;
ctl = RTA_DATA(tb[TCA_GRED_PARMS-1]);
if (ctl->DP > MAX_DPs-1 ) {
/* misbehaving is punished! Put in the default drop probability */
DPRINTK("\nGRED: DP %u not in the proper range fixed. New DP "
"set to default at %d\n",ctl->DP,table->def);
ctl->DP=table->def;
}
if (table->tab[ctl->DP] == NULL) {
table->tab[ctl->DP]=kmalloc(sizeof(struct gred_sched_data),
GFP_KERNEL);
if (NULL == table->tab[ctl->DP])
return -ENOMEM;
memset(table->tab[ctl->DP], 0, (sizeof(struct gred_sched_data)));
}
q= table->tab[ctl->DP];
if (table->grio) {
if (ctl->prio <=0) {
if (table->def && table->tab[table->def]) {
DPRINTK("\nGRED: DP %u does not have a prio"
"setting default to %d\n",ctl->DP,
table->tab[table->def]->prio);
q->prio=table->tab[table->def]->prio;
} else {
DPRINTK("\nGRED: DP %u does not have a prio"
" setting default to 8\n",ctl->DP);
q->prio=8;
}
} else {
q->prio=ctl->prio;
}
} else {
q->prio=8;
}
q->DP=ctl->DP;
q->Wlog = ctl->Wlog;
q->Plog = ctl->Plog;
q->limit = ctl->limit;
q->Scell_log = ctl->Scell_log;
q->Rmask = ctl->Plog < 32 ? ((1<<ctl->Plog) - 1) : ~0UL;
q->Scell_max = (255<<q->Scell_log);
q->qth_min = ctl->qth_min<<ctl->Wlog;
q->qth_max = ctl->qth_max<<ctl->Wlog;
q->qave=0;
q->backlog=0;
q->qcount = -1;
q->other=0;
q->forced=0;
q->pdrop=0;
q->early=0;
PSCHED_SET_PASTPERFECT(q->qidlestart);
memcpy(q->Stab, RTA_DATA(tb[TCA_GRED_STAB-1]), 256);
if ( table->initd && table->grio) {
/* this looks ugly but it's not in the fast path */
for (i=0;i<table->DPs;i++) {
if ((!table->tab[i]) || (i==q->DP) )
continue;
if (table->tab[i]->prio == q->prio ){
/* WRED mode detected */
table->eqp=1;
break;
}
}
}
if (!table->initd) {
table->initd=1;
/*
the first entry also goes into the default until
over-written
*/
if (table->tab[table->def] == NULL) {
table->tab[table->def]=
kmalloc(sizeof(struct gred_sched_data), GFP_KERNEL);
if (NULL == table->tab[table->def])
return -ENOMEM;
memset(table->tab[table->def], 0,
(sizeof(struct gred_sched_data)));
}
q= table->tab[table->def];
q->DP=table->def;
q->Wlog = ctl->Wlog;
q->Plog = ctl->Plog;
q->limit = ctl->limit;
q->Scell_log = ctl->Scell_log;
q->Rmask = ctl->Plog < 32 ? ((1<<ctl->Plog) - 1) : ~0UL;
q->Scell_max = (255<<q->Scell_log);
q->qth_min = ctl->qth_min<<ctl->Wlog;
q->qth_max = ctl->qth_max<<ctl->Wlog;
if (table->grio)
q->prio=table->tab[ctl->DP]->prio;
else
q->prio=8;
q->qcount = -1;
PSCHED_SET_PASTPERFECT(q->qidlestart);
memcpy(q->Stab, RTA_DATA(tb[TCA_GRED_STAB-1]), 256);
}
return 0;
}
static int gred_init(struct Qdisc *sch, struct rtattr *opt)
{
struct gred_sched *table = qdisc_priv(sch);
struct tc_gred_sopt *sopt;
struct rtattr *tb[TCA_GRED_STAB];
struct rtattr *tb2[TCA_GRED_DPS];
if (opt == NULL || rtattr_parse_nested(tb, TCA_GRED_STAB, opt))
return -EINVAL;
if (tb[TCA_GRED_PARMS-1] == 0 && tb[TCA_GRED_STAB-1] == 0) {
rtattr_parse_nested(tb2, TCA_GRED_DPS, opt);
if (tb2[TCA_GRED_DPS-1] == 0)
return -EINVAL;
sopt = RTA_DATA(tb2[TCA_GRED_DPS-1]);
table->DPs=sopt->DPs;
table->def=sopt->def_DP;
table->grio=sopt->grio;
table->initd=0;
return 0;
}
DPRINTK("\n GRED_INIT error!\n");
return -EINVAL;
}
static int gred_dump(struct Qdisc *sch, struct sk_buff *skb)
{
unsigned long qave;
struct rtattr *rta;
struct tc_gred_qopt *opt = NULL ;
struct tc_gred_qopt *dst;
struct gred_sched *table = qdisc_priv(sch);
struct gred_sched_data *q;
int i;
unsigned char *b = skb->tail;
rta = (struct rtattr*)b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
opt=kmalloc(sizeof(struct tc_gred_qopt)*MAX_DPs, GFP_KERNEL);
if (opt == NULL) {
DPRINTK("gred_dump:failed to malloc for %Zd\n",
sizeof(struct tc_gred_qopt)*MAX_DPs);
goto rtattr_failure;
}
memset(opt, 0, (sizeof(struct tc_gred_qopt))*table->DPs);
if (!table->initd) {
DPRINTK("NO GRED Queues setup!\n");
}
for (i=0;i<MAX_DPs;i++) {
dst= &opt[i];
q= table->tab[i];
if (!q) {
/* hack -- fix at some point with proper message
This is how we indicate to tc that there is no VQ
at this DP */
dst->DP=MAX_DPs+i;
continue;
}
dst->limit=q->limit;
dst->qth_min=q->qth_min>>q->Wlog;
dst->qth_max=q->qth_max>>q->Wlog;
dst->DP=q->DP;
dst->backlog=q->backlog;
if (q->qave) {
if (table->eqp && table->grio) {
q->qidlestart=table->tab[table->def]->qidlestart;
q->qave=table->tab[table->def]->qave;
}
if (!PSCHED_IS_PASTPERFECT(q->qidlestart)) {
long idle;
psched_time_t now;
PSCHED_GET_TIME(now);
idle = PSCHED_TDIFF_SAFE(now, q->qidlestart, q->Scell_max);
qave = q->qave >> q->Stab[(idle>>q->Scell_log)&0xFF];
dst->qave = qave >> q->Wlog;
} else {
dst->qave = q->qave >> q->Wlog;
}
} else {
dst->qave = 0;
}
dst->Wlog = q->Wlog;
dst->Plog = q->Plog;
dst->Scell_log = q->Scell_log;
dst->other = q->other;
dst->forced = q->forced;
dst->early = q->early;
dst->pdrop = q->pdrop;
dst->prio = q->prio;
dst->packets=q->packetsin;
dst->bytesin=q->bytesin;
}
RTA_PUT(skb, TCA_GRED_PARMS, sizeof(struct tc_gred_qopt)*MAX_DPs, opt);
rta->rta_len = skb->tail - b;
kfree(opt);
return skb->len;
rtattr_failure:
if (opt)
kfree(opt);
DPRINTK("gred_dump: FAILURE!!!!\n");
/* also free the opt struct here */
skb_trim(skb, b - skb->data);
return -1;
}
static void gred_destroy(struct Qdisc *sch)
{
struct gred_sched *table = qdisc_priv(sch);
int i;
for (i = 0;i < table->DPs; i++) {
if (table->tab[i])
kfree(table->tab[i]);
}
}
static struct Qdisc_ops gred_qdisc_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "gred",
.priv_size = sizeof(struct gred_sched),
.enqueue = gred_enqueue,
.dequeue = gred_dequeue,
.requeue = gred_requeue,
.drop = gred_drop,
.init = gred_init,
.reset = gred_reset,
.destroy = gred_destroy,
.change = gred_change,
.dump = gred_dump,
.owner = THIS_MODULE,
};
static int __init gred_module_init(void)
{
return register_qdisc(&gred_qdisc_ops);
}
static void __exit gred_module_exit(void)
{
unregister_qdisc(&gred_qdisc_ops);
}
module_init(gred_module_init)
module_exit(gred_module_exit)
MODULE_LICENSE("GPL");

1822
net/sched/sch_hfsc.c Normal file
View File

File diff suppressed because it is too large Load Diff

1759
net/sched/sch_htb.c Normal file
View File

File diff suppressed because it is too large Load Diff

436
net/sched/sch_ingress.c Normal file
View File

@@ -0,0 +1,436 @@
/* net/sched/sch_ingress.c - Ingress qdisc
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Jamal Hadi Salim 1999
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/types.h>
#include <linux/skbuff.h>
#include <linux/netdevice.h>
#include <linux/rtnetlink.h>
#include <linux/netfilter_ipv4.h>
#include <linux/netfilter_ipv6.h>
#include <linux/netfilter.h>
#include <linux/smp.h>
#include <net/pkt_sched.h>
#include <asm/byteorder.h>
#include <asm/uaccess.h>
#include <linux/kmod.h>
#include <linux/stat.h>
#include <linux/interrupt.h>
#include <linux/list.h>
#undef DEBUG_INGRESS
#ifdef DEBUG_INGRESS /* control */
#define DPRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define DPRINTK(format,args...)
#endif
#if 0 /* data */
#define D2PRINTK(format,args...) printk(KERN_DEBUG format,##args)
#else
#define D2PRINTK(format,args...)
#endif
#define PRIV(sch) qdisc_priv(sch)
/* Thanks to Doron Oz for this hack
*/
#ifndef CONFIG_NET_CLS_ACT
#ifdef CONFIG_NETFILTER
static int nf_registered;
#endif
#endif
struct ingress_qdisc_data {
struct Qdisc *q;
struct tcf_proto *filter_list;
};
/* ------------------------- Class/flow operations ------------------------- */
static int ingress_graft(struct Qdisc *sch,unsigned long arg,
struct Qdisc *new,struct Qdisc **old)
{
#ifdef DEBUG_INGRESS
struct ingress_qdisc_data *p = PRIV(sch);
#endif
DPRINTK("ingress_graft(sch %p,[qdisc %p],new %p,old %p)\n",
sch, p, new, old);
DPRINTK("\n ingress_graft: You cannot add qdiscs to classes");
return 1;
}
static struct Qdisc *ingress_leaf(struct Qdisc *sch, unsigned long arg)
{
return NULL;
}
static unsigned long ingress_get(struct Qdisc *sch,u32 classid)
{
#ifdef DEBUG_INGRESS
struct ingress_qdisc_data *p = PRIV(sch);
#endif
DPRINTK("ingress_get(sch %p,[qdisc %p],classid %x)\n", sch, p, classid);
return TC_H_MIN(classid) + 1;
}
static unsigned long ingress_bind_filter(struct Qdisc *sch,
unsigned long parent, u32 classid)
{
return ingress_get(sch, classid);
}
static void ingress_put(struct Qdisc *sch, unsigned long cl)
{
}
static int ingress_change(struct Qdisc *sch, u32 classid, u32 parent,
struct rtattr **tca, unsigned long *arg)
{
#ifdef DEBUG_INGRESS
struct ingress_qdisc_data *p = PRIV(sch);
#endif
DPRINTK("ingress_change(sch %p,[qdisc %p],classid %x,parent %x),"
"arg 0x%lx\n", sch, p, classid, parent, *arg);
DPRINTK("No effect. sch_ingress doesn't maintain classes at the moment");
return 0;
}
static void ingress_walk(struct Qdisc *sch,struct qdisc_walker *walker)
{
#ifdef DEBUG_INGRESS
struct ingress_qdisc_data *p = PRIV(sch);
#endif
DPRINTK("ingress_walk(sch %p,[qdisc %p],walker %p)\n", sch, p, walker);
DPRINTK("No effect. sch_ingress doesn't maintain classes at the moment");
}
static struct tcf_proto **ingress_find_tcf(struct Qdisc *sch,unsigned long cl)
{
struct ingress_qdisc_data *p = PRIV(sch);
return &p->filter_list;
}
/* --------------------------- Qdisc operations ---------------------------- */
static int ingress_enqueue(struct sk_buff *skb,struct Qdisc *sch)
{
struct ingress_qdisc_data *p = PRIV(sch);
struct tcf_result res;
int result;
D2PRINTK("ingress_enqueue(skb %p,sch %p,[qdisc %p])\n", skb, sch, p);
result = tc_classify(skb, p->filter_list, &res);
D2PRINTK("result %d class 0x%04x\n", result, res.classid);
/*
* Unlike normal "enqueue" functions, ingress_enqueue returns a
* firewall FW_* code.
*/
#ifdef CONFIG_NET_CLS_ACT
sch->bstats.packets++;
sch->bstats.bytes += skb->len;
switch (result) {
case TC_ACT_SHOT:
result = TC_ACT_SHOT;
sch->qstats.drops++;
break;
case TC_ACT_STOLEN:
case TC_ACT_QUEUED:
result = TC_ACT_STOLEN;
break;
case TC_ACT_RECLASSIFY:
case TC_ACT_OK:
case TC_ACT_UNSPEC:
default:
skb->tc_index = TC_H_MIN(res.classid);
result = TC_ACT_OK;
break;
};
/* backward compat */
#else
#ifdef CONFIG_NET_CLS_POLICE
switch (result) {
case TC_POLICE_SHOT:
result = NF_DROP;
sch->qstats.drops++;
break;
case TC_POLICE_RECLASSIFY: /* DSCP remarking here ? */
case TC_POLICE_OK:
case TC_POLICE_UNSPEC:
default:
sch->bstats.packets++;
sch->bstats.bytes += skb->len;
result = NF_ACCEPT;
break;
};
#else
D2PRINTK("Overriding result to ACCEPT\n");
result = NF_ACCEPT;
sch->bstats.packets++;
sch->bstats.bytes += skb->len;
#endif
#endif
return result;
}
static struct sk_buff *ingress_dequeue(struct Qdisc *sch)
{
/*
struct ingress_qdisc_data *p = PRIV(sch);
D2PRINTK("ingress_dequeue(sch %p,[qdisc %p])\n",sch,PRIV(p));
*/
return NULL;
}
static int ingress_requeue(struct sk_buff *skb,struct Qdisc *sch)
{
/*
struct ingress_qdisc_data *p = PRIV(sch);
D2PRINTK("ingress_requeue(skb %p,sch %p,[qdisc %p])\n",skb,sch,PRIV(p));
*/
return 0;
}
static unsigned int ingress_drop(struct Qdisc *sch)
{
#ifdef DEBUG_INGRESS
struct ingress_qdisc_data *p = PRIV(sch);
#endif
DPRINTK("ingress_drop(sch %p,[qdisc %p])\n", sch, p);
return 0;
}
#ifndef CONFIG_NET_CLS_ACT
#ifdef CONFIG_NETFILTER
static unsigned int
ing_hook(unsigned int hook, struct sk_buff **pskb,
const struct net_device *indev,
const struct net_device *outdev,
int (*okfn)(struct sk_buff *))
{
struct Qdisc *q;
struct sk_buff *skb = *pskb;
struct net_device *dev = skb->dev;
int fwres=NF_ACCEPT;
DPRINTK("ing_hook: skb %s dev=%s len=%u\n",
skb->sk ? "(owned)" : "(unowned)",
skb->dev ? (*pskb)->dev->name : "(no dev)",
skb->len);
/*
revisit later: Use a private since lock dev->queue_lock is also
used on the egress (might slow things for an iota)
*/
if (dev->qdisc_ingress) {
spin_lock(&dev->queue_lock);
if ((q = dev->qdisc_ingress) != NULL)
fwres = q->enqueue(skb, q);
spin_unlock(&dev->queue_lock);
}
return fwres;
}
/* after ipt_filter */
static struct nf_hook_ops ing_ops = {
.hook = ing_hook,
.owner = THIS_MODULE,
.pf = PF_INET,
.hooknum = NF_IP_PRE_ROUTING,
.priority = NF_IP_PRI_FILTER + 1,
};
static struct nf_hook_ops ing6_ops = {
.hook = ing_hook,
.owner = THIS_MODULE,
.pf = PF_INET6,
.hooknum = NF_IP6_PRE_ROUTING,
.priority = NF_IP6_PRI_FILTER + 1,
};
#endif
#endif
static int ingress_init(struct Qdisc *sch,struct rtattr *opt)
{
struct ingress_qdisc_data *p = PRIV(sch);
/* Make sure either netfilter or preferably CLS_ACT is
* compiled in */
#ifndef CONFIG_NET_CLS_ACT
#ifndef CONFIG_NETFILTER
printk("You MUST compile classifier actions into the kernel\n");
return -EINVAL;
#else
printk("Ingress scheduler: Classifier actions prefered over netfilter\n");
#endif
#endif
#ifndef CONFIG_NET_CLS_ACT
#ifdef CONFIG_NETFILTER
if (!nf_registered) {
if (nf_register_hook(&ing_ops) < 0) {
printk("ingress qdisc registration error \n");
return -EINVAL;
}
nf_registered++;
if (nf_register_hook(&ing6_ops) < 0) {
printk("IPv6 ingress qdisc registration error, " \
"disabling IPv6 support.\n");
} else
nf_registered++;
}
#endif
#endif
DPRINTK("ingress_init(sch %p,[qdisc %p],opt %p)\n",sch,p,opt);
p->q = &noop_qdisc;
return 0;
}
static void ingress_reset(struct Qdisc *sch)
{
struct ingress_qdisc_data *p = PRIV(sch);
DPRINTK("ingress_reset(sch %p,[qdisc %p])\n", sch, p);
/*
#if 0
*/
/* for future use */
qdisc_reset(p->q);
/*
#endif
*/
}
/* ------------------------------------------------------------- */
/* ------------------------------------------------------------- */
static void ingress_destroy(struct Qdisc *sch)
{
struct ingress_qdisc_data *p = PRIV(sch);
struct tcf_proto *tp;
DPRINTK("ingress_destroy(sch %p,[qdisc %p])\n", sch, p);
while (p->filter_list) {
tp = p->filter_list;
p->filter_list = tp->next;
tcf_destroy(tp);
}
#if 0
/* for future use */
qdisc_destroy(p->q);
#endif
}
static int ingress_dump(struct Qdisc *sch, struct sk_buff *skb)
{
unsigned char *b = skb->tail;
struct rtattr *rta;
rta = (struct rtattr *) b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
rta->rta_len = skb->tail - b;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct Qdisc_class_ops ingress_class_ops = {
.graft = ingress_graft,
.leaf = ingress_leaf,
.get = ingress_get,
.put = ingress_put,
.change = ingress_change,
.delete = NULL,
.walk = ingress_walk,
.tcf_chain = ingress_find_tcf,
.bind_tcf = ingress_bind_filter,
.unbind_tcf = ingress_put,
.dump = NULL,
};
static struct Qdisc_ops ingress_qdisc_ops = {
.next = NULL,
.cl_ops = &ingress_class_ops,
.id = "ingress",
.priv_size = sizeof(struct ingress_qdisc_data),
.enqueue = ingress_enqueue,
.dequeue = ingress_dequeue,
.requeue = ingress_requeue,
.drop = ingress_drop,
.init = ingress_init,
.reset = ingress_reset,
.destroy = ingress_destroy,
.change = NULL,
.dump = ingress_dump,
.owner = THIS_MODULE,
};
static int __init ingress_module_init(void)
{
int ret = 0;
if ((ret = register_qdisc(&ingress_qdisc_ops)) < 0) {
printk("Unable to register Ingress qdisc\n");
return ret;
}
return ret;
}
static void __exit ingress_module_exit(void)
{
unregister_qdisc(&ingress_qdisc_ops);
#ifndef CONFIG_NET_CLS_ACT
#ifdef CONFIG_NETFILTER
if (nf_registered) {
nf_unregister_hook(&ing_ops);
if (nf_registered > 1)
nf_unregister_hook(&ing6_ops);
}
#endif
#endif
}
module_init(ingress_module_init)
module_exit(ingress_module_exit)
MODULE_LICENSE("GPL");

598
net/sched/sch_netem.c Normal file
View File

@@ -0,0 +1,598 @@
/*
* net/sched/sch_netem.c Network emulator
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Many of the algorithms and ideas for this came from
* NIST Net which is not copyrighted.
*
* Authors: Stephen Hemminger <shemminger@osdl.org>
* Catalin(ux aka Dino) BOIE <catab at umbrella dot ro>
*/
#include <linux/config.h>
#include <linux/module.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <net/pkt_sched.h>
/* Network Emulation Queuing algorithm.
====================================
Sources: [1] Mark Carson, Darrin Santay, "NIST Net - A Linux-based
Network Emulation Tool
[2] Luigi Rizzo, DummyNet for FreeBSD
----------------------------------------------------------------
This started out as a simple way to delay outgoing packets to
test TCP but has grown to include most of the functionality
of a full blown network emulator like NISTnet. It can delay
packets and add random jitter (and correlation). The random
distribution can be loaded from a table as well to provide
normal, Pareto, or experimental curves. Packet loss,
duplication, and reordering can also be emulated.
This qdisc does not do classification that can be handled in
layering other disciplines. It does not need to do bandwidth
control either since that can be handled by using token
bucket or other rate control.
The simulator is limited by the Linux timer resolution
and will create packet bursts on the HZ boundary (1ms).
*/
struct netem_sched_data {
struct Qdisc *qdisc;
struct sk_buff_head delayed;
struct timer_list timer;
u32 latency;
u32 loss;
u32 limit;
u32 counter;
u32 gap;
u32 jitter;
u32 duplicate;
struct crndstate {
unsigned long last;
unsigned long rho;
} delay_cor, loss_cor, dup_cor;
struct disttable {
u32 size;
s16 table[0];
} *delay_dist;
};
/* Time stamp put into socket buffer control block */
struct netem_skb_cb {
psched_time_t time_to_send;
};
/* init_crandom - initialize correlated random number generator
* Use entropy source for initial seed.
*/
static void init_crandom(struct crndstate *state, unsigned long rho)
{
state->rho = rho;
state->last = net_random();
}
/* get_crandom - correlated random number generator
* Next number depends on last value.
* rho is scaled to avoid floating point.
*/
static unsigned long get_crandom(struct crndstate *state)
{
u64 value, rho;
unsigned long answer;
if (state->rho == 0) /* no correllation */
return net_random();
value = net_random();
rho = (u64)state->rho + 1;
answer = (value * ((1ull<<32) - rho) + state->last * rho) >> 32;
state->last = answer;
return answer;
}
/* tabledist - return a pseudo-randomly distributed value with mean mu and
* std deviation sigma. Uses table lookup to approximate the desired
* distribution, and a uniformly-distributed pseudo-random source.
*/
static long tabledist(unsigned long mu, long sigma,
struct crndstate *state, const struct disttable *dist)
{
long t, x;
unsigned long rnd;
if (sigma == 0)
return mu;
rnd = get_crandom(state);
/* default uniform distribution */
if (dist == NULL)
return (rnd % (2*sigma)) - sigma + mu;
t = dist->table[rnd % dist->size];
x = (sigma % NETEM_DIST_SCALE) * t;
if (x >= 0)
x += NETEM_DIST_SCALE/2;
else
x -= NETEM_DIST_SCALE/2;
return x / NETEM_DIST_SCALE + (sigma / NETEM_DIST_SCALE) * t + mu;
}
/* Put skb in the private delayed queue. */
static int delay_skb(struct Qdisc *sch, struct sk_buff *skb)
{
struct netem_sched_data *q = qdisc_priv(sch);
struct netem_skb_cb *cb = (struct netem_skb_cb *)skb->cb;
psched_tdiff_t td;
psched_time_t now;
PSCHED_GET_TIME(now);
td = tabledist(q->latency, q->jitter, &q->delay_cor, q->delay_dist);
PSCHED_TADD2(now, td, cb->time_to_send);
/* Always queue at tail to keep packets in order */
if (likely(q->delayed.qlen < q->limit)) {
__skb_queue_tail(&q->delayed, skb);
if (!timer_pending(&q->timer)) {
q->timer.expires = jiffies + PSCHED_US2JIFFIE(td);
add_timer(&q->timer);
}
return NET_XMIT_SUCCESS;
}
kfree_skb(skb);
return NET_XMIT_DROP;
}
static int netem_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct netem_sched_data *q = qdisc_priv(sch);
struct sk_buff *skb2;
int ret;
pr_debug("netem_enqueue skb=%p @%lu\n", skb, jiffies);
/* Random packet drop 0 => none, ~0 => all */
if (q->loss && q->loss >= get_crandom(&q->loss_cor)) {
pr_debug("netem_enqueue: random loss\n");
sch->qstats.drops++;
kfree_skb(skb);
return 0; /* lie about loss so TCP doesn't know */
}
/* Random duplication */
if (q->duplicate && q->duplicate >= get_crandom(&q->dup_cor)
&& (skb2 = skb_clone(skb, GFP_ATOMIC)) != NULL) {
pr_debug("netem_enqueue: dup %p\n", skb2);
if (delay_skb(sch, skb2)) {
sch->q.qlen++;
sch->bstats.bytes += skb2->len;
sch->bstats.packets++;
} else
sch->qstats.drops++;
}
/* If doing simple delay then gap == 0 so all packets
* go into the delayed holding queue
* otherwise if doing out of order only "1 out of gap"
* packets will be delayed.
*/
if (q->counter < q->gap) {
++q->counter;
ret = q->qdisc->enqueue(skb, q->qdisc);
} else {
q->counter = 0;
ret = delay_skb(sch, skb);
}
if (likely(ret == NET_XMIT_SUCCESS)) {
sch->q.qlen++;
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
} else
sch->qstats.drops++;
return ret;
}
/* Requeue packets but don't change time stamp */
static int netem_requeue(struct sk_buff *skb, struct Qdisc *sch)
{
struct netem_sched_data *q = qdisc_priv(sch);
int ret;
if ((ret = q->qdisc->ops->requeue(skb, q->qdisc)) == 0) {
sch->q.qlen++;
sch->qstats.requeues++;
}
return ret;
}
static unsigned int netem_drop(struct Qdisc* sch)
{
struct netem_sched_data *q = qdisc_priv(sch);
unsigned int len;
if ((len = q->qdisc->ops->drop(q->qdisc)) != 0) {
sch->q.qlen--;
sch->qstats.drops++;
}
return len;
}
/* Dequeue packet.
* Move all packets that are ready to send from the delay holding
* list to the underlying qdisc, then just call dequeue
*/
static struct sk_buff *netem_dequeue(struct Qdisc *sch)
{
struct netem_sched_data *q = qdisc_priv(sch);
struct sk_buff *skb;
skb = q->qdisc->dequeue(q->qdisc);
if (skb)
sch->q.qlen--;
return skb;
}
static void netem_watchdog(unsigned long arg)
{
struct Qdisc *sch = (struct Qdisc *)arg;
struct netem_sched_data *q = qdisc_priv(sch);
struct net_device *dev = sch->dev;
struct sk_buff *skb;
psched_time_t now;
pr_debug("netem_watchdog: fired @%lu\n", jiffies);
spin_lock_bh(&dev->queue_lock);
PSCHED_GET_TIME(now);
while ((skb = skb_peek(&q->delayed)) != NULL) {
const struct netem_skb_cb *cb
= (const struct netem_skb_cb *)skb->cb;
long delay
= PSCHED_US2JIFFIE(PSCHED_TDIFF(cb->time_to_send, now));
pr_debug("netem_watchdog: skb %p@%lu %ld\n",
skb, jiffies, delay);
/* if more time remaining? */
if (delay > 0) {
mod_timer(&q->timer, jiffies + delay);
break;
}
__skb_unlink(skb, &q->delayed);
if (q->qdisc->enqueue(skb, q->qdisc)) {
sch->q.qlen--;
sch->qstats.drops++;
}
}
qdisc_run(dev);
spin_unlock_bh(&dev->queue_lock);
}
static void netem_reset(struct Qdisc *sch)
{
struct netem_sched_data *q = qdisc_priv(sch);
qdisc_reset(q->qdisc);
skb_queue_purge(&q->delayed);
sch->q.qlen = 0;
del_timer_sync(&q->timer);
}
static int set_fifo_limit(struct Qdisc *q, int limit)
{
struct rtattr *rta;
int ret = -ENOMEM;
rta = kmalloc(RTA_LENGTH(sizeof(struct tc_fifo_qopt)), GFP_KERNEL);
if (rta) {
rta->rta_type = RTM_NEWQDISC;
rta->rta_len = RTA_LENGTH(sizeof(struct tc_fifo_qopt));
((struct tc_fifo_qopt *)RTA_DATA(rta))->limit = limit;
ret = q->ops->change(q, rta);
kfree(rta);
}
return ret;
}
/*
* Distribution data is a variable size payload containing
* signed 16 bit values.
*/
static int get_dist_table(struct Qdisc *sch, const struct rtattr *attr)
{
struct netem_sched_data *q = qdisc_priv(sch);
unsigned long n = RTA_PAYLOAD(attr)/sizeof(__s16);
const __s16 *data = RTA_DATA(attr);
struct disttable *d;
int i;
if (n > 65536)
return -EINVAL;
d = kmalloc(sizeof(*d) + n*sizeof(d->table[0]), GFP_KERNEL);
if (!d)
return -ENOMEM;
d->size = n;
for (i = 0; i < n; i++)
d->table[i] = data[i];
spin_lock_bh(&sch->dev->queue_lock);
d = xchg(&q->delay_dist, d);
spin_unlock_bh(&sch->dev->queue_lock);
kfree(d);
return 0;
}
static int get_correlation(struct Qdisc *sch, const struct rtattr *attr)
{
struct netem_sched_data *q = qdisc_priv(sch);
const struct tc_netem_corr *c = RTA_DATA(attr);
if (RTA_PAYLOAD(attr) != sizeof(*c))
return -EINVAL;
init_crandom(&q->delay_cor, c->delay_corr);
init_crandom(&q->loss_cor, c->loss_corr);
init_crandom(&q->dup_cor, c->dup_corr);
return 0;
}
static int netem_change(struct Qdisc *sch, struct rtattr *opt)
{
struct netem_sched_data *q = qdisc_priv(sch);
struct tc_netem_qopt *qopt;
int ret;
if (opt == NULL || RTA_PAYLOAD(opt) < sizeof(*qopt))
return -EINVAL;
qopt = RTA_DATA(opt);
ret = set_fifo_limit(q->qdisc, qopt->limit);
if (ret) {
pr_debug("netem: can't set fifo limit\n");
return ret;
}
q->latency = qopt->latency;
q->jitter = qopt->jitter;
q->limit = qopt->limit;
q->gap = qopt->gap;
q->loss = qopt->loss;
q->duplicate = qopt->duplicate;
/* Handle nested options after initial queue options.
* Should have put all options in nested format but too late now.
*/
if (RTA_PAYLOAD(opt) > sizeof(*qopt)) {
struct rtattr *tb[TCA_NETEM_MAX];
if (rtattr_parse(tb, TCA_NETEM_MAX,
RTA_DATA(opt) + sizeof(*qopt),
RTA_PAYLOAD(opt) - sizeof(*qopt)))
return -EINVAL;
if (tb[TCA_NETEM_CORR-1]) {
ret = get_correlation(sch, tb[TCA_NETEM_CORR-1]);
if (ret)
return ret;
}
if (tb[TCA_NETEM_DELAY_DIST-1]) {
ret = get_dist_table(sch, tb[TCA_NETEM_DELAY_DIST-1]);
if (ret)
return ret;
}
}
return 0;
}
static int netem_init(struct Qdisc *sch, struct rtattr *opt)
{
struct netem_sched_data *q = qdisc_priv(sch);
int ret;
if (!opt)
return -EINVAL;
skb_queue_head_init(&q->delayed);
init_timer(&q->timer);
q->timer.function = netem_watchdog;
q->timer.data = (unsigned long) sch;
q->counter = 0;
q->qdisc = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops);
if (!q->qdisc) {
pr_debug("netem: qdisc create failed\n");
return -ENOMEM;
}
ret = netem_change(sch, opt);
if (ret) {
pr_debug("netem: change failed\n");
qdisc_destroy(q->qdisc);
}
return ret;
}
static void netem_destroy(struct Qdisc *sch)
{
struct netem_sched_data *q = qdisc_priv(sch);
del_timer_sync(&q->timer);
qdisc_destroy(q->qdisc);
kfree(q->delay_dist);
}
static int netem_dump(struct Qdisc *sch, struct sk_buff *skb)
{
const struct netem_sched_data *q = qdisc_priv(sch);
unsigned char *b = skb->tail;
struct rtattr *rta = (struct rtattr *) b;
struct tc_netem_qopt qopt;
struct tc_netem_corr cor;
qopt.latency = q->latency;
qopt.jitter = q->jitter;
qopt.limit = q->limit;
qopt.loss = q->loss;
qopt.gap = q->gap;
qopt.duplicate = q->duplicate;
RTA_PUT(skb, TCA_OPTIONS, sizeof(qopt), &qopt);
cor.delay_corr = q->delay_cor.rho;
cor.loss_corr = q->loss_cor.rho;
cor.dup_corr = q->dup_cor.rho;
RTA_PUT(skb, TCA_NETEM_CORR, sizeof(cor), &cor);
rta->rta_len = skb->tail - b;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int netem_dump_class(struct Qdisc *sch, unsigned long cl,
struct sk_buff *skb, struct tcmsg *tcm)
{
struct netem_sched_data *q = qdisc_priv(sch);
if (cl != 1) /* only one class */
return -ENOENT;
tcm->tcm_handle |= TC_H_MIN(1);
tcm->tcm_info = q->qdisc->handle;
return 0;
}
static int netem_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
struct Qdisc **old)
{
struct netem_sched_data *q = qdisc_priv(sch);
if (new == NULL)
new = &noop_qdisc;
sch_tree_lock(sch);
*old = xchg(&q->qdisc, new);
qdisc_reset(*old);
sch->q.qlen = 0;
sch_tree_unlock(sch);
return 0;
}
static struct Qdisc *netem_leaf(struct Qdisc *sch, unsigned long arg)
{
struct netem_sched_data *q = qdisc_priv(sch);
return q->qdisc;
}
static unsigned long netem_get(struct Qdisc *sch, u32 classid)
{
return 1;
}
static void netem_put(struct Qdisc *sch, unsigned long arg)
{
}
static int netem_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
struct rtattr **tca, unsigned long *arg)
{
return -ENOSYS;
}
static int netem_delete(struct Qdisc *sch, unsigned long arg)
{
return -ENOSYS;
}
static void netem_walk(struct Qdisc *sch, struct qdisc_walker *walker)
{
if (!walker->stop) {
if (walker->count >= walker->skip)
if (walker->fn(sch, 1, walker) < 0) {
walker->stop = 1;
return;
}
walker->count++;
}
}
static struct tcf_proto **netem_find_tcf(struct Qdisc *sch, unsigned long cl)
{
return NULL;
}
static struct Qdisc_class_ops netem_class_ops = {
.graft = netem_graft,
.leaf = netem_leaf,
.get = netem_get,
.put = netem_put,
.change = netem_change_class,
.delete = netem_delete,
.walk = netem_walk,
.tcf_chain = netem_find_tcf,
.dump = netem_dump_class,
};
static struct Qdisc_ops netem_qdisc_ops = {
.id = "netem",
.cl_ops = &netem_class_ops,
.priv_size = sizeof(struct netem_sched_data),
.enqueue = netem_enqueue,
.dequeue = netem_dequeue,
.requeue = netem_requeue,
.drop = netem_drop,
.init = netem_init,
.reset = netem_reset,
.destroy = netem_destroy,
.change = netem_change,
.dump = netem_dump,
.owner = THIS_MODULE,
};
static int __init netem_module_init(void)
{
return register_qdisc(&netem_qdisc_ops);
}
static void __exit netem_module_exit(void)
{
unregister_qdisc(&netem_qdisc_ops);
}
module_init(netem_module_init)
module_exit(netem_module_exit)
MODULE_LICENSE("GPL");

444
net/sched/sch_prio.c Normal file
View File

@@ -0,0 +1,444 @@
/*
* net/sched/sch_prio.c Simple 3-band priority "scheduler".
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
* Fixes: 19990609: J Hadi Salim <hadi@nortelnetworks.com>:
* Init -- EINVAL when opt undefined
*/
#include <linux/config.h>
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
struct prio_sched_data
{
int bands;
struct tcf_proto *filter_list;
u8 prio2band[TC_PRIO_MAX+1];
struct Qdisc *queues[TCQ_PRIO_BANDS];
};
static struct Qdisc *
prio_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr)
{
struct prio_sched_data *q = qdisc_priv(sch);
u32 band = skb->priority;
struct tcf_result res;
*qerr = NET_XMIT_DROP;
if (TC_H_MAJ(skb->priority) != sch->handle) {
#ifdef CONFIG_NET_CLS_ACT
switch (tc_classify(skb, q->filter_list, &res)) {
case TC_ACT_STOLEN:
case TC_ACT_QUEUED:
*qerr = NET_XMIT_SUCCESS;
case TC_ACT_SHOT:
return NULL;
};
if (!q->filter_list ) {
#else
if (!q->filter_list || tc_classify(skb, q->filter_list, &res)) {
#endif
if (TC_H_MAJ(band))
band = 0;
return q->queues[q->prio2band[band&TC_PRIO_MAX]];
}
band = res.classid;
}
band = TC_H_MIN(band) - 1;
if (band > q->bands)
return q->queues[q->prio2band[0]];
return q->queues[band];
}
static int
prio_enqueue(struct sk_buff *skb, struct Qdisc *sch)
{
struct Qdisc *qdisc;
int ret;
qdisc = prio_classify(skb, sch, &ret);
#ifdef CONFIG_NET_CLS_ACT
if (qdisc == NULL) {
if (ret == NET_XMIT_DROP)
sch->qstats.drops++;
kfree_skb(skb);
return ret;
}
#endif
if ((ret = qdisc->enqueue(skb, qdisc)) == NET_XMIT_SUCCESS) {
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
sch->q.qlen++;
return NET_XMIT_SUCCESS;
}
sch->qstats.drops++;
return ret;
}
static int
prio_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
struct Qdisc *qdisc;
int ret;
qdisc = prio_classify(skb, sch, &ret);
#ifdef CONFIG_NET_CLS_ACT
if (qdisc == NULL) {
if (ret == NET_XMIT_DROP)
sch->qstats.drops++;
kfree_skb(skb);
return ret;
}
#endif
if ((ret = qdisc->ops->requeue(skb, qdisc)) == NET_XMIT_SUCCESS) {
sch->q.qlen++;
sch->qstats.requeues++;
return 0;
}
sch->qstats.drops++;
return NET_XMIT_DROP;
}
static struct sk_buff *
prio_dequeue(struct Qdisc* sch)
{
struct sk_buff *skb;
struct prio_sched_data *q = qdisc_priv(sch);
int prio;
struct Qdisc *qdisc;
for (prio = 0; prio < q->bands; prio++) {
qdisc = q->queues[prio];
skb = qdisc->dequeue(qdisc);
if (skb) {
sch->q.qlen--;
return skb;
}
}
return NULL;
}
static unsigned int prio_drop(struct Qdisc* sch)
{
struct prio_sched_data *q = qdisc_priv(sch);
int prio;
unsigned int len;
struct Qdisc *qdisc;
for (prio = q->bands-1; prio >= 0; prio--) {
qdisc = q->queues[prio];
if ((len = qdisc->ops->drop(qdisc)) != 0) {
sch->q.qlen--;
return len;
}
}
return 0;
}
static void
prio_reset(struct Qdisc* sch)
{
int prio;
struct prio_sched_data *q = qdisc_priv(sch);
for (prio=0; prio<q->bands; prio++)
qdisc_reset(q->queues[prio]);
sch->q.qlen = 0;
}
static void
prio_destroy(struct Qdisc* sch)
{
int prio;
struct prio_sched_data *q = qdisc_priv(sch);
struct tcf_proto *tp;
while ((tp = q->filter_list) != NULL) {
q->filter_list = tp->next;
tcf_destroy(tp);
}
for (prio=0; prio<q->bands; prio++)
qdisc_destroy(q->queues[prio]);
}
static int prio_tune(struct Qdisc *sch, struct rtattr *opt)
{
struct prio_sched_data *q = qdisc_priv(sch);
struct tc_prio_qopt *qopt = RTA_DATA(opt);
int i;
if (opt->rta_len < RTA_LENGTH(sizeof(*qopt)))
return -EINVAL;
if (qopt->bands > TCQ_PRIO_BANDS || qopt->bands < 2)
return -EINVAL;
for (i=0; i<=TC_PRIO_MAX; i++) {
if (qopt->priomap[i] >= qopt->bands)
return -EINVAL;
}
sch_tree_lock(sch);
q->bands = qopt->bands;
memcpy(q->prio2band, qopt->priomap, TC_PRIO_MAX+1);
for (i=q->bands; i<TCQ_PRIO_BANDS; i++) {
struct Qdisc *child = xchg(&q->queues[i], &noop_qdisc);
if (child != &noop_qdisc)
qdisc_destroy(child);
}
sch_tree_unlock(sch);
for (i=0; i<=TC_PRIO_MAX; i++) {
int band = q->prio2band[i];
if (q->queues[band] == &noop_qdisc) {
struct Qdisc *child;
child = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops);
if (child) {
sch_tree_lock(sch);
child = xchg(&q->queues[band], child);
if (child != &noop_qdisc)
qdisc_destroy(child);
sch_tree_unlock(sch);
}
}
}
return 0;
}
static int prio_init(struct Qdisc *sch, struct rtattr *opt)
{
struct prio_sched_data *q = qdisc_priv(sch);
int i;
for (i=0; i<TCQ_PRIO_BANDS; i++)
q->queues[i] = &noop_qdisc;
if (opt == NULL) {
return -EINVAL;
} else {
int err;
if ((err= prio_tune(sch, opt)) != 0)
return err;
}
return 0;
}
static int prio_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct prio_sched_data *q = qdisc_priv(sch);
unsigned char *b = skb->tail;
struct tc_prio_qopt opt;
opt.bands = q->bands;
memcpy(&opt.priomap, q->prio2band, TC_PRIO_MAX+1);
RTA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int prio_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
struct Qdisc **old)
{
struct prio_sched_data *q = qdisc_priv(sch);
unsigned long band = arg - 1;
if (band >= q->bands)
return -EINVAL;
if (new == NULL)
new = &noop_qdisc;
sch_tree_lock(sch);
*old = q->queues[band];
q->queues[band] = new;
sch->q.qlen -= (*old)->q.qlen;
qdisc_reset(*old);
sch_tree_unlock(sch);
return 0;
}
static struct Qdisc *
prio_leaf(struct Qdisc *sch, unsigned long arg)
{
struct prio_sched_data *q = qdisc_priv(sch);
unsigned long band = arg - 1;
if (band >= q->bands)
return NULL;
return q->queues[band];
}
static unsigned long prio_get(struct Qdisc *sch, u32 classid)
{
struct prio_sched_data *q = qdisc_priv(sch);
unsigned long band = TC_H_MIN(classid);
if (band - 1 >= q->bands)
return 0;
return band;
}
static unsigned long prio_bind(struct Qdisc *sch, unsigned long parent, u32 classid)
{
return prio_get(sch, classid);
}
static void prio_put(struct Qdisc *q, unsigned long cl)
{
return;
}
static int prio_change(struct Qdisc *sch, u32 handle, u32 parent, struct rtattr **tca, unsigned long *arg)
{
unsigned long cl = *arg;
struct prio_sched_data *q = qdisc_priv(sch);
if (cl - 1 > q->bands)
return -ENOENT;
return 0;
}
static int prio_delete(struct Qdisc *sch, unsigned long cl)
{
struct prio_sched_data *q = qdisc_priv(sch);
if (cl - 1 > q->bands)
return -ENOENT;
return 0;
}
static int prio_dump_class(struct Qdisc *sch, unsigned long cl, struct sk_buff *skb,
struct tcmsg *tcm)
{
struct prio_sched_data *q = qdisc_priv(sch);
if (cl - 1 > q->bands)
return -ENOENT;
tcm->tcm_handle |= TC_H_MIN(cl);
if (q->queues[cl-1])
tcm->tcm_info = q->queues[cl-1]->handle;
return 0;
}
static void prio_walk(struct Qdisc *sch, struct qdisc_walker *arg)
{
struct prio_sched_data *q = qdisc_priv(sch);
int prio;
if (arg->stop)
return;
for (prio = 0; prio < q->bands; prio++) {
if (arg->count < arg->skip) {
arg->count++;
continue;
}
if (arg->fn(sch, prio+1, arg) < 0) {
arg->stop = 1;
break;
}
arg->count++;
}
}
static struct tcf_proto ** prio_find_tcf(struct Qdisc *sch, unsigned long cl)
{
struct prio_sched_data *q = qdisc_priv(sch);
if (cl)
return NULL;
return &q->filter_list;
}
static struct Qdisc_class_ops prio_class_ops = {
.graft = prio_graft,
.leaf = prio_leaf,
.get = prio_get,
.put = prio_put,
.change = prio_change,
.delete = prio_delete,
.walk = prio_walk,
.tcf_chain = prio_find_tcf,
.bind_tcf = prio_bind,
.unbind_tcf = prio_put,
.dump = prio_dump_class,
};
static struct Qdisc_ops prio_qdisc_ops = {
.next = NULL,
.cl_ops = &prio_class_ops,
.id = "prio",
.priv_size = sizeof(struct prio_sched_data),
.enqueue = prio_enqueue,
.dequeue = prio_dequeue,
.requeue = prio_requeue,
.drop = prio_drop,
.init = prio_init,
.reset = prio_reset,
.destroy = prio_destroy,
.change = prio_tune,
.dump = prio_dump,
.owner = THIS_MODULE,
};
static int __init prio_module_init(void)
{
return register_qdisc(&prio_qdisc_ops);
}
static void __exit prio_module_exit(void)
{
unregister_qdisc(&prio_qdisc_ops);
}
module_init(prio_module_init)
module_exit(prio_module_exit)
MODULE_LICENSE("GPL");

459
net/sched/sch_red.c Normal file
View File

@@ -0,0 +1,459 @@
/*
* net/sched/sch_red.c Random Early Detection queue.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*
* Changes:
* J Hadi Salim <hadi@nortel.com> 980914: computation fixes
* Alexey Makarenko <makar@phoenix.kharkov.ua> 990814: qave on idle link was calculated incorrectly.
* J Hadi Salim <hadi@nortelnetworks.com> 980816: ECN support
*/
#include <linux/config.h>
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
#include <net/inet_ecn.h>
#include <net/dsfield.h>
/* Random Early Detection (RED) algorithm.
=======================================
Source: Sally Floyd and Van Jacobson, "Random Early Detection Gateways
for Congestion Avoidance", 1993, IEEE/ACM Transactions on Networking.
This file codes a "divisionless" version of RED algorithm
as written down in Fig.17 of the paper.
Short description.
------------------
When a new packet arrives we calculate the average queue length:
avg = (1-W)*avg + W*current_queue_len,
W is the filter time constant (chosen as 2^(-Wlog)), it controls
the inertia of the algorithm. To allow larger bursts, W should be
decreased.
if (avg > th_max) -> packet marked (dropped).
if (avg < th_min) -> packet passes.
if (th_min < avg < th_max) we calculate probability:
Pb = max_P * (avg - th_min)/(th_max-th_min)
and mark (drop) packet with this probability.
Pb changes from 0 (at avg==th_min) to max_P (avg==th_max).
max_P should be small (not 1), usually 0.01..0.02 is good value.
max_P is chosen as a number, so that max_P/(th_max-th_min)
is a negative power of two in order arithmetics to contain
only shifts.
Parameters, settable by user:
-----------------------------
limit - bytes (must be > qth_max + burst)
Hard limit on queue length, should be chosen >qth_max
to allow packet bursts. This parameter does not
affect the algorithms behaviour and can be chosen
arbitrarily high (well, less than ram size)
Really, this limit will never be reached
if RED works correctly.
qth_min - bytes (should be < qth_max/2)
qth_max - bytes (should be at least 2*qth_min and less limit)
Wlog - bits (<32) log(1/W).
Plog - bits (<32)
Plog is related to max_P by formula:
max_P = (qth_max-qth_min)/2^Plog;
F.e. if qth_max=128K and qth_min=32K, then Plog=22
corresponds to max_P=0.02
Scell_log
Stab
Lookup table for log((1-W)^(t/t_ave).
NOTES:
Upper bound on W.
-----------------
If you want to allow bursts of L packets of size S,
you should choose W:
L + 1 - th_min/S < (1-(1-W)^L)/W
th_min/S = 32 th_min/S = 4
log(W) L
-1 33
-2 35
-3 39
-4 46
-5 57
-6 75
-7 101
-8 135
-9 190
etc.
*/
struct red_sched_data
{
/* Parameters */
u32 limit; /* HARD maximal queue length */
u32 qth_min; /* Min average length threshold: A scaled */
u32 qth_max; /* Max average length threshold: A scaled */
u32 Rmask;
u32 Scell_max;
unsigned char flags;
char Wlog; /* log(W) */
char Plog; /* random number bits */
char Scell_log;
u8 Stab[256];
/* Variables */
unsigned long qave; /* Average queue length: A scaled */
int qcount; /* Packets since last random number generation */
u32 qR; /* Cached random number */
psched_time_t qidlestart; /* Start of idle period */
struct tc_red_xstats st;
};
static int red_ecn_mark(struct sk_buff *skb)
{
if (skb->nh.raw + 20 > skb->tail)
return 0;
switch (skb->protocol) {
case __constant_htons(ETH_P_IP):
if (INET_ECN_is_not_ect(skb->nh.iph->tos))
return 0;
IP_ECN_set_ce(skb->nh.iph);
return 1;
case __constant_htons(ETH_P_IPV6):
if (INET_ECN_is_not_ect(ipv6_get_dsfield(skb->nh.ipv6h)))
return 0;
IP6_ECN_set_ce(skb->nh.ipv6h);
return 1;
default:
return 0;
}
}
static int
red_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
psched_time_t now;
if (!PSCHED_IS_PASTPERFECT(q->qidlestart)) {
long us_idle;
int shift;
PSCHED_GET_TIME(now);
us_idle = PSCHED_TDIFF_SAFE(now, q->qidlestart, q->Scell_max);
PSCHED_SET_PASTPERFECT(q->qidlestart);
/*
The problem: ideally, average length queue recalcultion should
be done over constant clock intervals. This is too expensive, so that
the calculation is driven by outgoing packets.
When the queue is idle we have to model this clock by hand.
SF+VJ proposed to "generate" m = idletime/(average_pkt_size/bandwidth)
dummy packets as a burst after idle time, i.e.
q->qave *= (1-W)^m
This is an apparently overcomplicated solution (f.e. we have to precompute
a table to make this calculation in reasonable time)
I believe that a simpler model may be used here,
but it is field for experiments.
*/
shift = q->Stab[us_idle>>q->Scell_log];
if (shift) {
q->qave >>= shift;
} else {
/* Approximate initial part of exponent
with linear function:
(1-W)^m ~= 1-mW + ...
Seems, it is the best solution to
problem of too coarce exponent tabulation.
*/
us_idle = (q->qave * us_idle)>>q->Scell_log;
if (us_idle < q->qave/2)
q->qave -= us_idle;
else
q->qave >>= 1;
}
} else {
q->qave += sch->qstats.backlog - (q->qave >> q->Wlog);
/* NOTE:
q->qave is fixed point number with point at Wlog.
The formulae above is equvalent to floating point
version:
qave = qave*(1-W) + sch->qstats.backlog*W;
--ANK (980924)
*/
}
if (q->qave < q->qth_min) {
q->qcount = -1;
enqueue:
if (sch->qstats.backlog + skb->len <= q->limit) {
__skb_queue_tail(&sch->q, skb);
sch->qstats.backlog += skb->len;
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
return NET_XMIT_SUCCESS;
} else {
q->st.pdrop++;
}
kfree_skb(skb);
sch->qstats.drops++;
return NET_XMIT_DROP;
}
if (q->qave >= q->qth_max) {
q->qcount = -1;
sch->qstats.overlimits++;
mark:
if (!(q->flags&TC_RED_ECN) || !red_ecn_mark(skb)) {
q->st.early++;
goto drop;
}
q->st.marked++;
goto enqueue;
}
if (++q->qcount) {
/* The formula used below causes questions.
OK. qR is random number in the interval 0..Rmask
i.e. 0..(2^Plog). If we used floating point
arithmetics, it would be: (2^Plog)*rnd_num,
where rnd_num is less 1.
Taking into account, that qave have fixed
point at Wlog, and Plog is related to max_P by
max_P = (qth_max-qth_min)/2^Plog; two lines
below have the following floating point equivalent:
max_P*(qave - qth_min)/(qth_max-qth_min) < rnd/qcount
Any questions? --ANK (980924)
*/
if (((q->qave - q->qth_min)>>q->Wlog)*q->qcount < q->qR)
goto enqueue;
q->qcount = 0;
q->qR = net_random()&q->Rmask;
sch->qstats.overlimits++;
goto mark;
}
q->qR = net_random()&q->Rmask;
goto enqueue;
drop:
kfree_skb(skb);
sch->qstats.drops++;
return NET_XMIT_CN;
}
static int
red_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
PSCHED_SET_PASTPERFECT(q->qidlestart);
__skb_queue_head(&sch->q, skb);
sch->qstats.backlog += skb->len;
sch->qstats.requeues++;
return 0;
}
static struct sk_buff *
red_dequeue(struct Qdisc* sch)
{
struct sk_buff *skb;
struct red_sched_data *q = qdisc_priv(sch);
skb = __skb_dequeue(&sch->q);
if (skb) {
sch->qstats.backlog -= skb->len;
return skb;
}
PSCHED_GET_TIME(q->qidlestart);
return NULL;
}
static unsigned int red_drop(struct Qdisc* sch)
{
struct sk_buff *skb;
struct red_sched_data *q = qdisc_priv(sch);
skb = __skb_dequeue_tail(&sch->q);
if (skb) {
unsigned int len = skb->len;
sch->qstats.backlog -= len;
sch->qstats.drops++;
q->st.other++;
kfree_skb(skb);
return len;
}
PSCHED_GET_TIME(q->qidlestart);
return 0;
}
static void red_reset(struct Qdisc* sch)
{
struct red_sched_data *q = qdisc_priv(sch);
__skb_queue_purge(&sch->q);
sch->qstats.backlog = 0;
PSCHED_SET_PASTPERFECT(q->qidlestart);
q->qave = 0;
q->qcount = -1;
}
static int red_change(struct Qdisc *sch, struct rtattr *opt)
{
struct red_sched_data *q = qdisc_priv(sch);
struct rtattr *tb[TCA_RED_STAB];
struct tc_red_qopt *ctl;
if (opt == NULL ||
rtattr_parse_nested(tb, TCA_RED_STAB, opt) ||
tb[TCA_RED_PARMS-1] == 0 || tb[TCA_RED_STAB-1] == 0 ||
RTA_PAYLOAD(tb[TCA_RED_PARMS-1]) < sizeof(*ctl) ||
RTA_PAYLOAD(tb[TCA_RED_STAB-1]) < 256)
return -EINVAL;
ctl = RTA_DATA(tb[TCA_RED_PARMS-1]);
sch_tree_lock(sch);
q->flags = ctl->flags;
q->Wlog = ctl->Wlog;
q->Plog = ctl->Plog;
q->Rmask = ctl->Plog < 32 ? ((1<<ctl->Plog) - 1) : ~0UL;
q->Scell_log = ctl->Scell_log;
q->Scell_max = (255<<q->Scell_log);
q->qth_min = ctl->qth_min<<ctl->Wlog;
q->qth_max = ctl->qth_max<<ctl->Wlog;
q->limit = ctl->limit;
memcpy(q->Stab, RTA_DATA(tb[TCA_RED_STAB-1]), 256);
q->qcount = -1;
if (skb_queue_len(&sch->q) == 0)
PSCHED_SET_PASTPERFECT(q->qidlestart);
sch_tree_unlock(sch);
return 0;
}
static int red_init(struct Qdisc* sch, struct rtattr *opt)
{
return red_change(sch, opt);
}
static int red_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct red_sched_data *q = qdisc_priv(sch);
unsigned char *b = skb->tail;
struct rtattr *rta;
struct tc_red_qopt opt;
rta = (struct rtattr*)b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
opt.limit = q->limit;
opt.qth_min = q->qth_min>>q->Wlog;
opt.qth_max = q->qth_max>>q->Wlog;
opt.Wlog = q->Wlog;
opt.Plog = q->Plog;
opt.Scell_log = q->Scell_log;
opt.flags = q->flags;
RTA_PUT(skb, TCA_RED_PARMS, sizeof(opt), &opt);
rta->rta_len = skb->tail - b;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int red_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
{
struct red_sched_data *q = qdisc_priv(sch);
return gnet_stats_copy_app(d, &q->st, sizeof(q->st));
}
static struct Qdisc_ops red_qdisc_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "red",
.priv_size = sizeof(struct red_sched_data),
.enqueue = red_enqueue,
.dequeue = red_dequeue,
.requeue = red_requeue,
.drop = red_drop,
.init = red_init,
.reset = red_reset,
.change = red_change,
.dump = red_dump,
.dump_stats = red_dump_stats,
.owner = THIS_MODULE,
};
static int __init red_module_init(void)
{
return register_qdisc(&red_qdisc_ops);
}
static void __exit red_module_exit(void)
{
unregister_qdisc(&red_qdisc_ops);
}
module_init(red_module_init)
module_exit(red_module_exit)
MODULE_LICENSE("GPL");

497
net/sched/sch_sfq.c Normal file
View File

@@ -0,0 +1,497 @@
/*
* net/sched/sch_sfq.c Stochastic Fairness Queueing discipline.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
#include <linux/config.h>
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/jiffies.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <linux/init.h>
#include <net/ip.h>
#include <linux/ipv6.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
/* Stochastic Fairness Queuing algorithm.
=======================================
Source:
Paul E. McKenney "Stochastic Fairness Queuing",
IEEE INFOCOMM'90 Proceedings, San Francisco, 1990.
Paul E. McKenney "Stochastic Fairness Queuing",
"Interworking: Research and Experience", v.2, 1991, p.113-131.
See also:
M. Shreedhar and George Varghese "Efficient Fair
Queuing using Deficit Round Robin", Proc. SIGCOMM 95.
This is not the thing that is usually called (W)FQ nowadays.
It does not use any timestamp mechanism, but instead
processes queues in round-robin order.
ADVANTAGE:
- It is very cheap. Both CPU and memory requirements are minimal.
DRAWBACKS:
- "Stochastic" -> It is not 100% fair.
When hash collisions occur, several flows are considered as one.
- "Round-robin" -> It introduces larger delays than virtual clock
based schemes, and should not be used for isolating interactive
traffic from non-interactive. It means, that this scheduler
should be used as leaf of CBQ or P3, which put interactive traffic
to higher priority band.
We still need true WFQ for top level CSZ, but using WFQ
for the best effort traffic is absolutely pointless:
SFQ is superior for this purpose.
IMPLEMENTATION:
This implementation limits maximal queue length to 128;
maximal mtu to 2^15-1; number of hash buckets to 1024.
The only goal of this restrictions was that all data
fit into one 4K page :-). Struct sfq_sched_data is
organized in anti-cache manner: all the data for a bucket
are scattered over different locations. This is not good,
but it allowed me to put it into 4K.
It is easy to increase these values, but not in flight. */
#define SFQ_DEPTH 128
#define SFQ_HASH_DIVISOR 1024
/* This type should contain at least SFQ_DEPTH*2 values */
typedef unsigned char sfq_index;
struct sfq_head
{
sfq_index next;
sfq_index prev;
};
struct sfq_sched_data
{
/* Parameters */
int perturb_period;
unsigned quantum; /* Allotment per round: MUST BE >= MTU */
int limit;
/* Variables */
struct timer_list perturb_timer;
int perturbation;
sfq_index tail; /* Index of current slot in round */
sfq_index max_depth; /* Maximal depth */
sfq_index ht[SFQ_HASH_DIVISOR]; /* Hash table */
sfq_index next[SFQ_DEPTH]; /* Active slots link */
short allot[SFQ_DEPTH]; /* Current allotment per slot */
unsigned short hash[SFQ_DEPTH]; /* Hash value indexed by slots */
struct sk_buff_head qs[SFQ_DEPTH]; /* Slot queue */
struct sfq_head dep[SFQ_DEPTH*2]; /* Linked list of slots, indexed by depth */
};
static __inline__ unsigned sfq_fold_hash(struct sfq_sched_data *q, u32 h, u32 h1)
{
int pert = q->perturbation;
/* Have we any rotation primitives? If not, WHY? */
h ^= (h1<<pert) ^ (h1>>(0x1F - pert));
h ^= h>>10;
return h & 0x3FF;
}
static unsigned sfq_hash(struct sfq_sched_data *q, struct sk_buff *skb)
{
u32 h, h2;
switch (skb->protocol) {
case __constant_htons(ETH_P_IP):
{
struct iphdr *iph = skb->nh.iph;
h = iph->daddr;
h2 = iph->saddr^iph->protocol;
if (!(iph->frag_off&htons(IP_MF|IP_OFFSET)) &&
(iph->protocol == IPPROTO_TCP ||
iph->protocol == IPPROTO_UDP ||
iph->protocol == IPPROTO_ESP))
h2 ^= *(((u32*)iph) + iph->ihl);
break;
}
case __constant_htons(ETH_P_IPV6):
{
struct ipv6hdr *iph = skb->nh.ipv6h;
h = iph->daddr.s6_addr32[3];
h2 = iph->saddr.s6_addr32[3]^iph->nexthdr;
if (iph->nexthdr == IPPROTO_TCP ||
iph->nexthdr == IPPROTO_UDP ||
iph->nexthdr == IPPROTO_ESP)
h2 ^= *(u32*)&iph[1];
break;
}
default:
h = (u32)(unsigned long)skb->dst^skb->protocol;
h2 = (u32)(unsigned long)skb->sk;
}
return sfq_fold_hash(q, h, h2);
}
static inline void sfq_link(struct sfq_sched_data *q, sfq_index x)
{
sfq_index p, n;
int d = q->qs[x].qlen + SFQ_DEPTH;
p = d;
n = q->dep[d].next;
q->dep[x].next = n;
q->dep[x].prev = p;
q->dep[p].next = q->dep[n].prev = x;
}
static inline void sfq_dec(struct sfq_sched_data *q, sfq_index x)
{
sfq_index p, n;
n = q->dep[x].next;
p = q->dep[x].prev;
q->dep[p].next = n;
q->dep[n].prev = p;
if (n == p && q->max_depth == q->qs[x].qlen + 1)
q->max_depth--;
sfq_link(q, x);
}
static inline void sfq_inc(struct sfq_sched_data *q, sfq_index x)
{
sfq_index p, n;
int d;
n = q->dep[x].next;
p = q->dep[x].prev;
q->dep[p].next = n;
q->dep[n].prev = p;
d = q->qs[x].qlen;
if (q->max_depth < d)
q->max_depth = d;
sfq_link(q, x);
}
static unsigned int sfq_drop(struct Qdisc *sch)
{
struct sfq_sched_data *q = qdisc_priv(sch);
sfq_index d = q->max_depth;
struct sk_buff *skb;
unsigned int len;
/* Queue is full! Find the longest slot and
drop a packet from it */
if (d > 1) {
sfq_index x = q->dep[d+SFQ_DEPTH].next;
skb = q->qs[x].prev;
len = skb->len;
__skb_unlink(skb, &q->qs[x]);
kfree_skb(skb);
sfq_dec(q, x);
sch->q.qlen--;
sch->qstats.drops++;
return len;
}
if (d == 1) {
/* It is difficult to believe, but ALL THE SLOTS HAVE LENGTH 1. */
d = q->next[q->tail];
q->next[q->tail] = q->next[d];
q->allot[q->next[d]] += q->quantum;
skb = q->qs[d].prev;
len = skb->len;
__skb_unlink(skb, &q->qs[d]);
kfree_skb(skb);
sfq_dec(q, d);
sch->q.qlen--;
q->ht[q->hash[d]] = SFQ_DEPTH;
sch->qstats.drops++;
return len;
}
return 0;
}
static int
sfq_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct sfq_sched_data *q = qdisc_priv(sch);
unsigned hash = sfq_hash(q, skb);
sfq_index x;
x = q->ht[hash];
if (x == SFQ_DEPTH) {
q->ht[hash] = x = q->dep[SFQ_DEPTH].next;
q->hash[x] = hash;
}
__skb_queue_tail(&q->qs[x], skb);
sfq_inc(q, x);
if (q->qs[x].qlen == 1) { /* The flow is new */
if (q->tail == SFQ_DEPTH) { /* It is the first flow */
q->tail = x;
q->next[x] = x;
q->allot[x] = q->quantum;
} else {
q->next[x] = q->next[q->tail];
q->next[q->tail] = x;
q->tail = x;
}
}
if (++sch->q.qlen < q->limit-1) {
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
return 0;
}
sfq_drop(sch);
return NET_XMIT_CN;
}
static int
sfq_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
struct sfq_sched_data *q = qdisc_priv(sch);
unsigned hash = sfq_hash(q, skb);
sfq_index x;
x = q->ht[hash];
if (x == SFQ_DEPTH) {
q->ht[hash] = x = q->dep[SFQ_DEPTH].next;
q->hash[x] = hash;
}
__skb_queue_head(&q->qs[x], skb);
sfq_inc(q, x);
if (q->qs[x].qlen == 1) { /* The flow is new */
if (q->tail == SFQ_DEPTH) { /* It is the first flow */
q->tail = x;
q->next[x] = x;
q->allot[x] = q->quantum;
} else {
q->next[x] = q->next[q->tail];
q->next[q->tail] = x;
q->tail = x;
}
}
if (++sch->q.qlen < q->limit - 1) {
sch->qstats.requeues++;
return 0;
}
sch->qstats.drops++;
sfq_drop(sch);
return NET_XMIT_CN;
}
static struct sk_buff *
sfq_dequeue(struct Qdisc* sch)
{
struct sfq_sched_data *q = qdisc_priv(sch);
struct sk_buff *skb;
sfq_index a, old_a;
/* No active slots */
if (q->tail == SFQ_DEPTH)
return NULL;
a = old_a = q->next[q->tail];
/* Grab packet */
skb = __skb_dequeue(&q->qs[a]);
sfq_dec(q, a);
sch->q.qlen--;
/* Is the slot empty? */
if (q->qs[a].qlen == 0) {
q->ht[q->hash[a]] = SFQ_DEPTH;
a = q->next[a];
if (a == old_a) {
q->tail = SFQ_DEPTH;
return skb;
}
q->next[q->tail] = a;
q->allot[a] += q->quantum;
} else if ((q->allot[a] -= skb->len) <= 0) {
q->tail = a;
a = q->next[a];
q->allot[a] += q->quantum;
}
return skb;
}
static void
sfq_reset(struct Qdisc* sch)
{
struct sk_buff *skb;
while ((skb = sfq_dequeue(sch)) != NULL)
kfree_skb(skb);
}
static void sfq_perturbation(unsigned long arg)
{
struct Qdisc *sch = (struct Qdisc*)arg;
struct sfq_sched_data *q = qdisc_priv(sch);
q->perturbation = net_random()&0x1F;
if (q->perturb_period) {
q->perturb_timer.expires = jiffies + q->perturb_period;
add_timer(&q->perturb_timer);
}
}
static int sfq_change(struct Qdisc *sch, struct rtattr *opt)
{
struct sfq_sched_data *q = qdisc_priv(sch);
struct tc_sfq_qopt *ctl = RTA_DATA(opt);
if (opt->rta_len < RTA_LENGTH(sizeof(*ctl)))
return -EINVAL;
sch_tree_lock(sch);
q->quantum = ctl->quantum ? : psched_mtu(sch->dev);
q->perturb_period = ctl->perturb_period*HZ;
if (ctl->limit)
q->limit = min_t(u32, ctl->limit, SFQ_DEPTH);
while (sch->q.qlen >= q->limit-1)
sfq_drop(sch);
del_timer(&q->perturb_timer);
if (q->perturb_period) {
q->perturb_timer.expires = jiffies + q->perturb_period;
add_timer(&q->perturb_timer);
}
sch_tree_unlock(sch);
return 0;
}
static int sfq_init(struct Qdisc *sch, struct rtattr *opt)
{
struct sfq_sched_data *q = qdisc_priv(sch);
int i;
init_timer(&q->perturb_timer);
q->perturb_timer.data = (unsigned long)sch;
q->perturb_timer.function = sfq_perturbation;
for (i=0; i<SFQ_HASH_DIVISOR; i++)
q->ht[i] = SFQ_DEPTH;
for (i=0; i<SFQ_DEPTH; i++) {
skb_queue_head_init(&q->qs[i]);
q->dep[i+SFQ_DEPTH].next = i+SFQ_DEPTH;
q->dep[i+SFQ_DEPTH].prev = i+SFQ_DEPTH;
}
q->limit = SFQ_DEPTH;
q->max_depth = 0;
q->tail = SFQ_DEPTH;
if (opt == NULL) {
q->quantum = psched_mtu(sch->dev);
q->perturb_period = 0;
} else {
int err = sfq_change(sch, opt);
if (err)
return err;
}
for (i=0; i<SFQ_DEPTH; i++)
sfq_link(q, i);
return 0;
}
static void sfq_destroy(struct Qdisc *sch)
{
struct sfq_sched_data *q = qdisc_priv(sch);
del_timer(&q->perturb_timer);
}
static int sfq_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct sfq_sched_data *q = qdisc_priv(sch);
unsigned char *b = skb->tail;
struct tc_sfq_qopt opt;
opt.quantum = q->quantum;
opt.perturb_period = q->perturb_period/HZ;
opt.limit = q->limit;
opt.divisor = SFQ_HASH_DIVISOR;
opt.flows = q->limit;
RTA_PUT(skb, TCA_OPTIONS, sizeof(opt), &opt);
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static struct Qdisc_ops sfq_qdisc_ops = {
.next = NULL,
.cl_ops = NULL,
.id = "sfq",
.priv_size = sizeof(struct sfq_sched_data),
.enqueue = sfq_enqueue,
.dequeue = sfq_dequeue,
.requeue = sfq_requeue,
.drop = sfq_drop,
.init = sfq_init,
.reset = sfq_reset,
.destroy = sfq_destroy,
.change = NULL,
.dump = sfq_dump,
.owner = THIS_MODULE,
};
static int __init sfq_module_init(void)
{
return register_qdisc(&sfq_qdisc_ops);
}
static void __exit sfq_module_exit(void)
{
unregister_qdisc(&sfq_qdisc_ops);
}
module_init(sfq_module_init)
module_exit(sfq_module_exit)
MODULE_LICENSE("GPL");

543
net/sched/sch_tbf.c Normal file
View File

@@ -0,0 +1,543 @@
/*
* net/sched/sch_tbf.c Token Bucket Filter queue.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
* Dmitry Torokhov <dtor@mail.ru> - allow attaching inner qdiscs -
* original idea by Martin Devera
*
*/
#include <linux/config.h>
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/jiffies.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
/* Simple Token Bucket Filter.
=======================================
SOURCE.
-------
None.
Description.
------------
A data flow obeys TBF with rate R and depth B, if for any
time interval t_i...t_f the number of transmitted bits
does not exceed B + R*(t_f-t_i).
Packetized version of this definition:
The sequence of packets of sizes s_i served at moments t_i
obeys TBF, if for any i<=k:
s_i+....+s_k <= B + R*(t_k - t_i)
Algorithm.
----------
Let N(t_i) be B/R initially and N(t) grow continuously with time as:
N(t+delta) = min{B/R, N(t) + delta}
If the first packet in queue has length S, it may be
transmitted only at the time t_* when S/R <= N(t_*),
and in this case N(t) jumps:
N(t_* + 0) = N(t_* - 0) - S/R.
Actually, QoS requires two TBF to be applied to a data stream.
One of them controls steady state burst size, another
one with rate P (peak rate) and depth M (equal to link MTU)
limits bursts at a smaller time scale.
It is easy to see that P>R, and B>M. If P is infinity, this double
TBF is equivalent to a single one.
When TBF works in reshaping mode, latency is estimated as:
lat = max ((L-B)/R, (L-M)/P)
NOTES.
------
If TBF throttles, it starts a watchdog timer, which will wake it up
when it is ready to transmit.
Note that the minimal timer resolution is 1/HZ.
If no new packets arrive during this period,
or if the device is not awaken by EOI for some previous packet,
TBF can stop its activity for 1/HZ.
This means, that with depth B, the maximal rate is
R_crit = B*HZ
F.e. for 10Mbit ethernet and HZ=100 the minimal allowed B is ~10Kbytes.
Note that the peak rate TBF is much more tough: with MTU 1500
P_crit = 150Kbytes/sec. So, if you need greater peak
rates, use alpha with HZ=1000 :-)
With classful TBF, limit is just kept for backwards compatibility.
It is passed to the default bfifo qdisc - if the inner qdisc is
changed the limit is not effective anymore.
*/
struct tbf_sched_data
{
/* Parameters */
u32 limit; /* Maximal length of backlog: bytes */
u32 buffer; /* Token bucket depth/rate: MUST BE >= MTU/B */
u32 mtu;
u32 max_size;
struct qdisc_rate_table *R_tab;
struct qdisc_rate_table *P_tab;
/* Variables */
long tokens; /* Current number of B tokens */
long ptokens; /* Current number of P tokens */
psched_time_t t_c; /* Time check-point */
struct timer_list wd_timer; /* Watchdog timer */
struct Qdisc *qdisc; /* Inner qdisc, default - bfifo queue */
};
#define L2T(q,L) ((q)->R_tab->data[(L)>>(q)->R_tab->rate.cell_log])
#define L2T_P(q,L) ((q)->P_tab->data[(L)>>(q)->P_tab->rate.cell_log])
static int tbf_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct tbf_sched_data *q = qdisc_priv(sch);
int ret;
if (skb->len > q->max_size) {
sch->qstats.drops++;
#ifdef CONFIG_NET_CLS_POLICE
if (sch->reshape_fail == NULL || sch->reshape_fail(skb, sch))
#endif
kfree_skb(skb);
return NET_XMIT_DROP;
}
if ((ret = q->qdisc->enqueue(skb, q->qdisc)) != 0) {
sch->qstats.drops++;
return ret;
}
sch->q.qlen++;
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
return 0;
}
static int tbf_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
struct tbf_sched_data *q = qdisc_priv(sch);
int ret;
if ((ret = q->qdisc->ops->requeue(skb, q->qdisc)) == 0) {
sch->q.qlen++;
sch->qstats.requeues++;
}
return ret;
}
static unsigned int tbf_drop(struct Qdisc* sch)
{
struct tbf_sched_data *q = qdisc_priv(sch);
unsigned int len;
if ((len = q->qdisc->ops->drop(q->qdisc)) != 0) {
sch->q.qlen--;
sch->qstats.drops++;
}
return len;
}
static void tbf_watchdog(unsigned long arg)
{
struct Qdisc *sch = (struct Qdisc*)arg;
sch->flags &= ~TCQ_F_THROTTLED;
netif_schedule(sch->dev);
}
static struct sk_buff *tbf_dequeue(struct Qdisc* sch)
{
struct tbf_sched_data *q = qdisc_priv(sch);
struct sk_buff *skb;
skb = q->qdisc->dequeue(q->qdisc);
if (skb) {
psched_time_t now;
long toks, delay;
long ptoks = 0;
unsigned int len = skb->len;
PSCHED_GET_TIME(now);
toks = PSCHED_TDIFF_SAFE(now, q->t_c, q->buffer);
if (q->P_tab) {
ptoks = toks + q->ptokens;
if (ptoks > (long)q->mtu)
ptoks = q->mtu;
ptoks -= L2T_P(q, len);
}
toks += q->tokens;
if (toks > (long)q->buffer)
toks = q->buffer;
toks -= L2T(q, len);
if ((toks|ptoks) >= 0) {
q->t_c = now;
q->tokens = toks;
q->ptokens = ptoks;
sch->q.qlen--;
sch->flags &= ~TCQ_F_THROTTLED;
return skb;
}
delay = PSCHED_US2JIFFIE(max_t(long, -toks, -ptoks));
if (delay == 0)
delay = 1;
mod_timer(&q->wd_timer, jiffies+delay);
/* Maybe we have a shorter packet in the queue,
which can be sent now. It sounds cool,
but, however, this is wrong in principle.
We MUST NOT reorder packets under these circumstances.
Really, if we split the flow into independent
subflows, it would be a very good solution.
This is the main idea of all FQ algorithms
(cf. CSZ, HPFQ, HFSC)
*/
if (q->qdisc->ops->requeue(skb, q->qdisc) != NET_XMIT_SUCCESS) {
/* When requeue fails skb is dropped */
sch->q.qlen--;
sch->qstats.drops++;
}
sch->flags |= TCQ_F_THROTTLED;
sch->qstats.overlimits++;
}
return NULL;
}
static void tbf_reset(struct Qdisc* sch)
{
struct tbf_sched_data *q = qdisc_priv(sch);
qdisc_reset(q->qdisc);
sch->q.qlen = 0;
PSCHED_GET_TIME(q->t_c);
q->tokens = q->buffer;
q->ptokens = q->mtu;
sch->flags &= ~TCQ_F_THROTTLED;
del_timer(&q->wd_timer);
}
static struct Qdisc *tbf_create_dflt_qdisc(struct net_device *dev, u32 limit)
{
struct Qdisc *q = qdisc_create_dflt(dev, &bfifo_qdisc_ops);
struct rtattr *rta;
int ret;
if (q) {
rta = kmalloc(RTA_LENGTH(sizeof(struct tc_fifo_qopt)), GFP_KERNEL);
if (rta) {
rta->rta_type = RTM_NEWQDISC;
rta->rta_len = RTA_LENGTH(sizeof(struct tc_fifo_qopt));
((struct tc_fifo_qopt *)RTA_DATA(rta))->limit = limit;
ret = q->ops->change(q, rta);
kfree(rta);
if (ret == 0)
return q;
}
qdisc_destroy(q);
}
return NULL;
}
static int tbf_change(struct Qdisc* sch, struct rtattr *opt)
{
int err = -EINVAL;
struct tbf_sched_data *q = qdisc_priv(sch);
struct rtattr *tb[TCA_TBF_PTAB];
struct tc_tbf_qopt *qopt;
struct qdisc_rate_table *rtab = NULL;
struct qdisc_rate_table *ptab = NULL;
struct Qdisc *child = NULL;
int max_size,n;
if (rtattr_parse_nested(tb, TCA_TBF_PTAB, opt) ||
tb[TCA_TBF_PARMS-1] == NULL ||
RTA_PAYLOAD(tb[TCA_TBF_PARMS-1]) < sizeof(*qopt))
goto done;
qopt = RTA_DATA(tb[TCA_TBF_PARMS-1]);
rtab = qdisc_get_rtab(&qopt->rate, tb[TCA_TBF_RTAB-1]);
if (rtab == NULL)
goto done;
if (qopt->peakrate.rate) {
if (qopt->peakrate.rate > qopt->rate.rate)
ptab = qdisc_get_rtab(&qopt->peakrate, tb[TCA_TBF_PTAB-1]);
if (ptab == NULL)
goto done;
}
for (n = 0; n < 256; n++)
if (rtab->data[n] > qopt->buffer) break;
max_size = (n << qopt->rate.cell_log)-1;
if (ptab) {
int size;
for (n = 0; n < 256; n++)
if (ptab->data[n] > qopt->mtu) break;
size = (n << qopt->peakrate.cell_log)-1;
if (size < max_size) max_size = size;
}
if (max_size < 0)
goto done;
if (q->qdisc == &noop_qdisc) {
if ((child = tbf_create_dflt_qdisc(sch->dev, qopt->limit)) == NULL)
goto done;
}
sch_tree_lock(sch);
if (child) q->qdisc = child;
q->limit = qopt->limit;
q->mtu = qopt->mtu;
q->max_size = max_size;
q->buffer = qopt->buffer;
q->tokens = q->buffer;
q->ptokens = q->mtu;
rtab = xchg(&q->R_tab, rtab);
ptab = xchg(&q->P_tab, ptab);
sch_tree_unlock(sch);
err = 0;
done:
if (rtab)
qdisc_put_rtab(rtab);
if (ptab)
qdisc_put_rtab(ptab);
return err;
}
static int tbf_init(struct Qdisc* sch, struct rtattr *opt)
{
struct tbf_sched_data *q = qdisc_priv(sch);
if (opt == NULL)
return -EINVAL;
PSCHED_GET_TIME(q->t_c);
init_timer(&q->wd_timer);
q->wd_timer.function = tbf_watchdog;
q->wd_timer.data = (unsigned long)sch;
q->qdisc = &noop_qdisc;
return tbf_change(sch, opt);
}
static void tbf_destroy(struct Qdisc *sch)
{
struct tbf_sched_data *q = qdisc_priv(sch);
del_timer(&q->wd_timer);
if (q->P_tab)
qdisc_put_rtab(q->P_tab);
if (q->R_tab)
qdisc_put_rtab(q->R_tab);
qdisc_destroy(q->qdisc);
}
static int tbf_dump(struct Qdisc *sch, struct sk_buff *skb)
{
struct tbf_sched_data *q = qdisc_priv(sch);
unsigned char *b = skb->tail;
struct rtattr *rta;
struct tc_tbf_qopt opt;
rta = (struct rtattr*)b;
RTA_PUT(skb, TCA_OPTIONS, 0, NULL);
opt.limit = q->limit;
opt.rate = q->R_tab->rate;
if (q->P_tab)
opt.peakrate = q->P_tab->rate;
else
memset(&opt.peakrate, 0, sizeof(opt.peakrate));
opt.mtu = q->mtu;
opt.buffer = q->buffer;
RTA_PUT(skb, TCA_TBF_PARMS, sizeof(opt), &opt);
rta->rta_len = skb->tail - b;
return skb->len;
rtattr_failure:
skb_trim(skb, b - skb->data);
return -1;
}
static int tbf_dump_class(struct Qdisc *sch, unsigned long cl,
struct sk_buff *skb, struct tcmsg *tcm)
{
struct tbf_sched_data *q = qdisc_priv(sch);
if (cl != 1) /* only one class */
return -ENOENT;
tcm->tcm_handle |= TC_H_MIN(1);
tcm->tcm_info = q->qdisc->handle;
return 0;
}
static int tbf_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new,
struct Qdisc **old)
{
struct tbf_sched_data *q = qdisc_priv(sch);
if (new == NULL)
new = &noop_qdisc;
sch_tree_lock(sch);
*old = xchg(&q->qdisc, new);
qdisc_reset(*old);
sch->q.qlen = 0;
sch_tree_unlock(sch);
return 0;
}
static struct Qdisc *tbf_leaf(struct Qdisc *sch, unsigned long arg)
{
struct tbf_sched_data *q = qdisc_priv(sch);
return q->qdisc;
}
static unsigned long tbf_get(struct Qdisc *sch, u32 classid)
{
return 1;
}
static void tbf_put(struct Qdisc *sch, unsigned long arg)
{
}
static int tbf_change_class(struct Qdisc *sch, u32 classid, u32 parentid,
struct rtattr **tca, unsigned long *arg)
{
return -ENOSYS;
}
static int tbf_delete(struct Qdisc *sch, unsigned long arg)
{
return -ENOSYS;
}
static void tbf_walk(struct Qdisc *sch, struct qdisc_walker *walker)
{
if (!walker->stop) {
if (walker->count >= walker->skip)
if (walker->fn(sch, 1, walker) < 0) {
walker->stop = 1;
return;
}
walker->count++;
}
}
static struct tcf_proto **tbf_find_tcf(struct Qdisc *sch, unsigned long cl)
{
return NULL;
}
static struct Qdisc_class_ops tbf_class_ops =
{
.graft = tbf_graft,
.leaf = tbf_leaf,
.get = tbf_get,
.put = tbf_put,
.change = tbf_change_class,
.delete = tbf_delete,
.walk = tbf_walk,
.tcf_chain = tbf_find_tcf,
.dump = tbf_dump_class,
};
static struct Qdisc_ops tbf_qdisc_ops = {
.next = NULL,
.cl_ops = &tbf_class_ops,
.id = "tbf",
.priv_size = sizeof(struct tbf_sched_data),
.enqueue = tbf_enqueue,
.dequeue = tbf_dequeue,
.requeue = tbf_requeue,
.drop = tbf_drop,
.init = tbf_init,
.reset = tbf_reset,
.destroy = tbf_destroy,
.change = tbf_change,
.dump = tbf_dump,
.owner = THIS_MODULE,
};
static int __init tbf_module_init(void)
{
return register_qdisc(&tbf_qdisc_ops);
}
static void __exit tbf_module_exit(void)
{
unregister_qdisc(&tbf_qdisc_ops);
}
module_init(tbf_module_init)
module_exit(tbf_module_exit)
MODULE_LICENSE("GPL");

511
net/sched/sch_teql.c Normal file
View File

@@ -0,0 +1,511 @@
/* net/sched/sch_teql.c "True" (or "trivial") link equalizer.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
*/
#include <linux/module.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <linux/bitops.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/string.h>
#include <linux/mm.h>
#include <linux/socket.h>
#include <linux/sockios.h>
#include <linux/in.h>
#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/if_ether.h>
#include <linux/inet.h>
#include <linux/netdevice.h>
#include <linux/etherdevice.h>
#include <linux/notifier.h>
#include <linux/init.h>
#include <net/ip.h>
#include <net/route.h>
#include <linux/skbuff.h>
#include <linux/moduleparam.h>
#include <net/sock.h>
#include <net/pkt_sched.h>
/*
How to setup it.
----------------
After loading this module you will find a new device teqlN
and new qdisc with the same name. To join a slave to the equalizer
you should just set this qdisc on a device f.e.
# tc qdisc add dev eth0 root teql0
# tc qdisc add dev eth1 root teql0
That's all. Full PnP 8)
Applicability.
--------------
1. Slave devices MUST be active devices, i.e., they must raise the tbusy
signal and generate EOI events. If you want to equalize virtual devices
like tunnels, use a normal eql device.
2. This device puts no limitations on physical slave characteristics
f.e. it will equalize 9600baud line and 100Mb ethernet perfectly :-)
Certainly, large difference in link speeds will make the resulting
eqalized link unusable, because of huge packet reordering.
I estimate an upper useful difference as ~10 times.
3. If the slave requires address resolution, only protocols using
neighbour cache (IPv4/IPv6) will work over the equalized link.
Other protocols are still allowed to use the slave device directly,
which will not break load balancing, though native slave
traffic will have the highest priority. */
struct teql_master
{
struct Qdisc_ops qops;
struct net_device *dev;
struct Qdisc *slaves;
struct list_head master_list;
struct net_device_stats stats;
};
struct teql_sched_data
{
struct Qdisc *next;
struct teql_master *m;
struct neighbour *ncache;
struct sk_buff_head q;
};
#define NEXT_SLAVE(q) (((struct teql_sched_data*)qdisc_priv(q))->next)
#define FMASK (IFF_BROADCAST|IFF_POINTOPOINT|IFF_BROADCAST)
/* "teql*" qdisc routines */
static int
teql_enqueue(struct sk_buff *skb, struct Qdisc* sch)
{
struct net_device *dev = sch->dev;
struct teql_sched_data *q = qdisc_priv(sch);
__skb_queue_tail(&q->q, skb);
if (q->q.qlen <= dev->tx_queue_len) {
sch->bstats.bytes += skb->len;
sch->bstats.packets++;
return 0;
}
__skb_unlink(skb, &q->q);
kfree_skb(skb);
sch->qstats.drops++;
return NET_XMIT_DROP;
}
static int
teql_requeue(struct sk_buff *skb, struct Qdisc* sch)
{
struct teql_sched_data *q = qdisc_priv(sch);
__skb_queue_head(&q->q, skb);
sch->qstats.requeues++;
return 0;
}
static struct sk_buff *
teql_dequeue(struct Qdisc* sch)
{
struct teql_sched_data *dat = qdisc_priv(sch);
struct sk_buff *skb;
skb = __skb_dequeue(&dat->q);
if (skb == NULL) {
struct net_device *m = dat->m->dev->qdisc->dev;
if (m) {
dat->m->slaves = sch;
netif_wake_queue(m);
}
}
sch->q.qlen = dat->q.qlen + dat->m->dev->qdisc->q.qlen;
return skb;
}
static __inline__ void
teql_neigh_release(struct neighbour *n)
{
if (n)
neigh_release(n);
}
static void
teql_reset(struct Qdisc* sch)
{
struct teql_sched_data *dat = qdisc_priv(sch);
skb_queue_purge(&dat->q);
sch->q.qlen = 0;
teql_neigh_release(xchg(&dat->ncache, NULL));
}
static void
teql_destroy(struct Qdisc* sch)
{
struct Qdisc *q, *prev;
struct teql_sched_data *dat = qdisc_priv(sch);
struct teql_master *master = dat->m;
if ((prev = master->slaves) != NULL) {
do {
q = NEXT_SLAVE(prev);
if (q == sch) {
NEXT_SLAVE(prev) = NEXT_SLAVE(q);
if (q == master->slaves) {
master->slaves = NEXT_SLAVE(q);
if (q == master->slaves) {
master->slaves = NULL;
spin_lock_bh(&master->dev->queue_lock);
qdisc_reset(master->dev->qdisc);
spin_unlock_bh(&master->dev->queue_lock);
}
}
skb_queue_purge(&dat->q);
teql_neigh_release(xchg(&dat->ncache, NULL));
break;
}
} while ((prev = q) != master->slaves);
}
}
static int teql_qdisc_init(struct Qdisc *sch, struct rtattr *opt)
{
struct net_device *dev = sch->dev;
struct teql_master *m = (struct teql_master*)sch->ops;
struct teql_sched_data *q = qdisc_priv(sch);
if (dev->hard_header_len > m->dev->hard_header_len)
return -EINVAL;
if (m->dev == dev)
return -ELOOP;
q->m = m;
skb_queue_head_init(&q->q);
if (m->slaves) {
if (m->dev->flags & IFF_UP) {
if ((m->dev->flags&IFF_POINTOPOINT && !(dev->flags&IFF_POINTOPOINT))
|| (m->dev->flags&IFF_BROADCAST && !(dev->flags&IFF_BROADCAST))
|| (m->dev->flags&IFF_MULTICAST && !(dev->flags&IFF_MULTICAST))
|| dev->mtu < m->dev->mtu)
return -EINVAL;
} else {
if (!(dev->flags&IFF_POINTOPOINT))
m->dev->flags &= ~IFF_POINTOPOINT;
if (!(dev->flags&IFF_BROADCAST))
m->dev->flags &= ~IFF_BROADCAST;
if (!(dev->flags&IFF_MULTICAST))
m->dev->flags &= ~IFF_MULTICAST;
if (dev->mtu < m->dev->mtu)
m->dev->mtu = dev->mtu;
}
q->next = NEXT_SLAVE(m->slaves);
NEXT_SLAVE(m->slaves) = sch;
} else {
q->next = sch;
m->slaves = sch;
m->dev->mtu = dev->mtu;
m->dev->flags = (m->dev->flags&~FMASK)|(dev->flags&FMASK);
}
return 0;
}
/* "teql*" netdevice routines */
static int
__teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev)
{
struct teql_sched_data *q = qdisc_priv(dev->qdisc);
struct neighbour *mn = skb->dst->neighbour;
struct neighbour *n = q->ncache;
if (mn->tbl == NULL)
return -EINVAL;
if (n && n->tbl == mn->tbl &&
memcmp(n->primary_key, mn->primary_key, mn->tbl->key_len) == 0) {
atomic_inc(&n->refcnt);
} else {
n = __neigh_lookup_errno(mn->tbl, mn->primary_key, dev);
if (IS_ERR(n))
return PTR_ERR(n);
}
if (neigh_event_send(n, skb_res) == 0) {
int err;
read_lock(&n->lock);
err = dev->hard_header(skb, dev, ntohs(skb->protocol), n->ha, NULL, skb->len);
read_unlock(&n->lock);
if (err < 0) {
neigh_release(n);
return -EINVAL;
}
teql_neigh_release(xchg(&q->ncache, n));
return 0;
}
neigh_release(n);
return (skb_res == NULL) ? -EAGAIN : 1;
}
static __inline__ int
teql_resolve(struct sk_buff *skb, struct sk_buff *skb_res, struct net_device *dev)
{
if (dev->hard_header == NULL ||
skb->dst == NULL ||
skb->dst->neighbour == NULL)
return 0;
return __teql_resolve(skb, skb_res, dev);
}
static int teql_master_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct teql_master *master = (void*)dev->priv;
struct Qdisc *start, *q;
int busy;
int nores;
int len = skb->len;
struct sk_buff *skb_res = NULL;
start = master->slaves;
restart:
nores = 0;
busy = 0;
if ((q = start) == NULL)
goto drop;
do {
struct net_device *slave = q->dev;
if (slave->qdisc_sleeping != q)
continue;
if (netif_queue_stopped(slave) || ! netif_running(slave)) {
busy = 1;
continue;
}
switch (teql_resolve(skb, skb_res, slave)) {
case 0:
if (spin_trylock(&slave->xmit_lock)) {
slave->xmit_lock_owner = smp_processor_id();
if (!netif_queue_stopped(slave) &&
slave->hard_start_xmit(skb, slave) == 0) {
slave->xmit_lock_owner = -1;
spin_unlock(&slave->xmit_lock);
master->slaves = NEXT_SLAVE(q);
netif_wake_queue(dev);
master->stats.tx_packets++;
master->stats.tx_bytes += len;
return 0;
}
slave->xmit_lock_owner = -1;
spin_unlock(&slave->xmit_lock);
}
if (netif_queue_stopped(dev))
busy = 1;
break;
case 1:
master->slaves = NEXT_SLAVE(q);
return 0;
default:
nores = 1;
break;
}
__skb_pull(skb, skb->nh.raw - skb->data);
} while ((q = NEXT_SLAVE(q)) != start);
if (nores && skb_res == NULL) {
skb_res = skb;
goto restart;
}
if (busy) {
netif_stop_queue(dev);
return 1;
}
master->stats.tx_errors++;
drop:
master->stats.tx_dropped++;
dev_kfree_skb(skb);
return 0;
}
static int teql_master_open(struct net_device *dev)
{
struct Qdisc * q;
struct teql_master *m = (void*)dev->priv;
int mtu = 0xFFFE;
unsigned flags = IFF_NOARP|IFF_MULTICAST;
if (m->slaves == NULL)
return -EUNATCH;
flags = FMASK;
q = m->slaves;
do {
struct net_device *slave = q->dev;
if (slave == NULL)
return -EUNATCH;
if (slave->mtu < mtu)
mtu = slave->mtu;
if (slave->hard_header_len > LL_MAX_HEADER)
return -EINVAL;
/* If all the slaves are BROADCAST, master is BROADCAST
If all the slaves are PtP, master is PtP
Otherwise, master is NBMA.
*/
if (!(slave->flags&IFF_POINTOPOINT))
flags &= ~IFF_POINTOPOINT;
if (!(slave->flags&IFF_BROADCAST))
flags &= ~IFF_BROADCAST;
if (!(slave->flags&IFF_MULTICAST))
flags &= ~IFF_MULTICAST;
} while ((q = NEXT_SLAVE(q)) != m->slaves);
m->dev->mtu = mtu;
m->dev->flags = (m->dev->flags&~FMASK) | flags;
netif_start_queue(m->dev);
return 0;
}
static int teql_master_close(struct net_device *dev)
{
netif_stop_queue(dev);
return 0;
}
static struct net_device_stats *teql_master_stats(struct net_device *dev)
{
struct teql_master *m = (void*)dev->priv;
return &m->stats;
}
static int teql_master_mtu(struct net_device *dev, int new_mtu)
{
struct teql_master *m = (void*)dev->priv;
struct Qdisc *q;
if (new_mtu < 68)
return -EINVAL;
q = m->slaves;
if (q) {
do {
if (new_mtu > q->dev->mtu)
return -EINVAL;
} while ((q=NEXT_SLAVE(q)) != m->slaves);
}
dev->mtu = new_mtu;
return 0;
}
static __init void teql_master_setup(struct net_device *dev)
{
struct teql_master *master = dev->priv;
struct Qdisc_ops *ops = &master->qops;
master->dev = dev;
ops->priv_size = sizeof(struct teql_sched_data);
ops->enqueue = teql_enqueue;
ops->dequeue = teql_dequeue;
ops->requeue = teql_requeue;
ops->init = teql_qdisc_init;
ops->reset = teql_reset;
ops->destroy = teql_destroy;
ops->owner = THIS_MODULE;
dev->open = teql_master_open;
dev->hard_start_xmit = teql_master_xmit;
dev->stop = teql_master_close;
dev->get_stats = teql_master_stats;
dev->change_mtu = teql_master_mtu;
dev->type = ARPHRD_VOID;
dev->mtu = 1500;
dev->tx_queue_len = 100;
dev->flags = IFF_NOARP;
dev->hard_header_len = LL_MAX_HEADER;
SET_MODULE_OWNER(dev);
}
static LIST_HEAD(master_dev_list);
static int max_equalizers = 1;
module_param(max_equalizers, int, 0);
MODULE_PARM_DESC(max_equalizers, "Max number of link equalizers");
static int __init teql_init(void)
{
int i;
int err = -ENODEV;
for (i = 0; i < max_equalizers; i++) {
struct net_device *dev;
struct teql_master *master;
dev = alloc_netdev(sizeof(struct teql_master),
"teql%d", teql_master_setup);
if (!dev) {
err = -ENOMEM;
break;
}
if ((err = register_netdev(dev))) {
free_netdev(dev);
break;
}
master = dev->priv;
strlcpy(master->qops.id, dev->name, IFNAMSIZ);
err = register_qdisc(&master->qops);
if (err) {
unregister_netdev(dev);
free_netdev(dev);
break;
}
list_add_tail(&master->master_list, &master_dev_list);
}
return i ? 0 : err;
}
static void __exit teql_exit(void)
{
struct teql_master *master, *nxt;
list_for_each_entry_safe(master, nxt, &master_dev_list, master_list) {
list_del(&master->master_list);
unregister_qdisc(&master->qops);
unregister_netdev(master->dev);
free_netdev(master->dev);
}
}
module_init(teql_init);
module_exit(teql_exit);
MODULE_LICENSE("GPL");