updated iproute2 to 2.6.23 + latest debian patches + latest esfq. deleted patches have been merged upstream. note, this also brings in the necessary esfq changes in the kernel

SVN-Revision: 10147
This commit is contained in:
Tim Yardley 2008-01-08 22:28:35 +00:00
parent 244e25e2d7
commit 063e150eb9
6 changed files with 3267 additions and 1772 deletions

View file

@ -9,14 +9,14 @@
include $(TOPDIR)/rules.mk
PKG_NAME:=iproute2
PKG_VERSION:=2.6.20-070313
PKG_VERSION:=2.6.23
PKG_RELEASE:=1
PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.gz
PKG_SOURCE:=$(PKG_NAME)-$(PKG_VERSION).tar.bz2
PKG_SOURCE_URL:=http://developer.osdl.org/dev/iproute2/download/
PKG_MD5SUM:=7bc5883aadf740761fa2dd70b661e8cc
PKG_MD5SUM:=
PKG_BUILD_DIR:=$(BUILD_DIR)/iproute-$(PKG_VERSION)
PKG_BUILD_DIR:=$(BUILD_DIR)/iproute2-$(PKG_VERSION)
include $(INCLUDE_DIR)/package.mk

File diff suppressed because it is too large Load diff

View file

@ -1,9 +0,0 @@
Index: iproute-2.6.20-070313/Config
===================================================================
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ iproute-2.6.20-070313/Config 2007-06-09 13:53:58.000000000 +0100
@@ -0,0 +1,4 @@
+# Fixed config to disable ATM support even if present on host system
+TC_CONFIG_ATM:=n
+TC_CONFIG_ACTION_GACT=y
+TC_CONFIG_ACTION_PROB=y

View file

@ -1,98 +0,0 @@
Index: iproute-2.6.20-070313/tc/q_htb.c
===================================================================
--- iproute-2.6.20-070313.orig/tc/q_htb.c 2007-06-09 13:53:57.000000000 +0100
+++ iproute-2.6.20-070313/tc/q_htb.c 2007-06-09 13:54:00.000000000 +0100
@@ -35,10 +35,14 @@
" default minor id of class to which unclassified packets are sent {0}\n"
" r2q DRR quantums are computed as rate in Bps/r2q {10}\n"
" debug string of 16 numbers each 0-3 {0}\n\n"
- "... class add ... htb rate R1 burst B1 [prio P] [slot S] [pslot PS]\n"
+ "... class add ... htb rate R1 [burst B1] [mpu B] [overhead O]\n"
+ " [prio P] [slot S] [pslot PS]\n"
" [ceil R2] [cburst B2] [mtu MTU] [quantum Q]\n"
" rate rate allocated to this class (class can still borrow)\n"
" burst max bytes burst which can be accumulated during idle period {computed}\n"
+ " mpu minimum packet size used in rate computations\n"
+ " overhead per-packet size overhead used in rate computations\n"
+
" ceil definite upper class rate (no borrows) {rate}\n"
" cburst burst but for ceil {computed}\n"
" mtu max packet size we create rate map for {1600}\n"
@@ -103,7 +107,9 @@
struct tc_htb_opt opt;
__u32 rtab[256],ctab[256];
unsigned buffer=0,cbuffer=0;
- int cell_log=-1,ccell_log = -1,mtu;
+ int cell_log=-1,ccell_log = -1;
+ unsigned mtu, mpu;
+ unsigned char mpu8 = 0, overhead = 0;
struct rtattr *tail;
memset(&opt, 0, sizeof(opt)); mtu = 1600; /* eth packet len */
@@ -120,6 +126,16 @@
if (get_u32(&mtu, *argv, 10)) {
explain1("mtu"); return -1;
}
+ } else if (matches(*argv, "mpu") == 0) {
+ NEXT_ARG();
+ if (get_u8(&mpu8, *argv, 10)) {
+ explain1("mpu"); return -1;
+ }
+ } else if (matches(*argv, "overhead") == 0) {
+ NEXT_ARG();
+ if (get_u8(&overhead, *argv, 10)) {
+ explain1("overhead"); return -1;
+ }
} else if (matches(*argv, "quantum") == 0) {
NEXT_ARG();
if (get_u32(&opt.quantum, *argv, 10)) {
@@ -191,14 +207,18 @@
if (!buffer) buffer = opt.rate.rate / HZ + mtu;
if (!cbuffer) cbuffer = opt.ceil.rate / HZ + mtu;
- if ((cell_log = tc_calc_rtable(opt.rate.rate, rtab, cell_log, mtu, 0)) < 0) {
+/* encode overhead and mpu, 8 bits each, into lower 16 bits */
+ mpu = (unsigned)mpu8 | (unsigned)overhead << 8;
+ opt.ceil.mpu = mpu; opt.rate.mpu = mpu;
+
+ if ((cell_log = tc_calc_rtable(opt.rate.rate, rtab, cell_log, mtu, mpu)) < 0) {
fprintf(stderr, "htb: failed to calculate rate table.\n");
return -1;
}
opt.buffer = tc_calc_xmittime(opt.rate.rate, buffer);
opt.rate.cell_log = cell_log;
- if ((ccell_log = tc_calc_rtable(opt.ceil.rate, ctab, cell_log, mtu, 0)) < 0) {
+ if ((ccell_log = tc_calc_rtable(opt.ceil.rate, ctab, cell_log, mtu, mpu)) < 0) {
fprintf(stderr, "htb: failed to calculate ceil rate table.\n");
return -1;
}
@@ -222,6 +242,7 @@
double buffer,cbuffer;
SPRINT_BUF(b1);
SPRINT_BUF(b2);
+ SPRINT_BUF(b3);
if (opt == NULL)
return 0;
@@ -244,10 +265,16 @@
fprintf(f, "ceil %s ", sprint_rate(hopt->ceil.rate, b1));
cbuffer = ((double)hopt->ceil.rate*tc_core_tick2usec(hopt->cbuffer))/1000000;
if (show_details) {
- fprintf(f, "burst %s/%u mpu %s ", sprint_size(buffer, b1),
- 1<<hopt->rate.cell_log, sprint_size(hopt->rate.mpu, b2));
- fprintf(f, "cburst %s/%u mpu %s ", sprint_size(cbuffer, b1),
- 1<<hopt->ceil.cell_log, sprint_size(hopt->ceil.mpu, b2));
+ fprintf(f, "burst %s/%u mpu %s overhead %s ",
+ sprint_size(buffer, b1),
+ 1<<hopt->rate.cell_log,
+ sprint_size(hopt->rate.mpu&0xFF, b2),
+ sprint_size((hopt->rate.mpu>>8)&0xFF, b3));
+ fprintf(f, "cburst %s/%u mpu %s overhead %s ",
+ sprint_size(cbuffer, b1),
+ 1<<hopt->ceil.cell_log,
+ sprint_size(hopt->ceil.mpu&0xFF, b2),
+ sprint_size((hopt->ceil.mpu>>8)&0xFF, b3));
fprintf(f, "level %d ", (int)hopt->level);
} else {
fprintf(f, "burst %s ", sprint_size(buffer, b1));

View file

@ -1,12 +1,12 @@
diff -urN --exclude=.svn iproute2-2.6.11-050330/include/linux/pkt_sched.h iproute2-2.6.11-050330/include/linux/pkt_sched.h
--- iproute2-2.6.11-050330/include/linux/pkt_sched.h 2007-05-04 22:21:48.000000000 -0400
+++ iproute2-2.6.11-050330/include/linux/pkt_sched.h 2007-05-04 22:27:12.000000000 -0400
@@ -174,8 +174,38 @@
diff -Naur iproute-2.6.20-070313.orig/include/linux/pkt_sched.h iproute-2.6.20-070313/include/linux/pkt_sched.h
--- iproute-2.6.20-070313.orig/include/linux/pkt_sched.h 2007-03-13 14:50:56.000000000 -0700
+++ iproute-2.6.20-070313/include/linux/pkt_sched.h 2007-06-09 11:32:22.000000000 -0700
@@ -146,8 +146,37 @@
*
* The only reason for this is efficiency, it is possible
* to change these parameters in compile time.
+ *
+ * If you need to play with these values use esfq instead.
+ * If you need to play with these values, use esfq instead.
*/
+/* ESFQ section */
@ -17,6 +17,7 @@ diff -urN --exclude=.svn iproute2-2.6.11-050330/include/linux/pkt_sched.h iprout
+ TCA_SFQ_HASH_CLASSIC,
+ TCA_SFQ_HASH_DST,
+ TCA_SFQ_HASH_SRC,
+ TCA_SFQ_HASH_FWMARK,
+ /* conntrack */
+ TCA_SFQ_HASH_CTORIGDST,
+ TCA_SFQ_HASH_CTORIGSRC,
@ -34,54 +35,14 @@ diff -urN --exclude=.svn iproute2-2.6.11-050330/include/linux/pkt_sched.h iprout
+ unsigned flows; /* Maximal number of flows */
+ unsigned hash_kind; /* Hash function to use for flow identification */
+};
+
+
+
/* RED section */
enum
@@ -551,8 +580,37 @@
*
* The only reason for this is efficiency, it is possible
* to change these parameters in compile time.
+ *
+ * If you need to play with these values use esfq instead.
*/
+/* ESFQ section */
+
+enum
+{
+ /* traditional */
+ TCA_SFQ_HASH_CLASSIC,
+ TCA_SFQ_HASH_DST,
+ TCA_SFQ_HASH_SRC,
+ /* conntrack */
+ TCA_SFQ_HASH_CTORIGDST,
+ TCA_SFQ_HASH_CTORIGSRC,
+ TCA_SFQ_HASH_CTREPLDST,
+ TCA_SFQ_HASH_CTREPLSRC,
+ TCA_SFQ_HASH_CTNATCHG,
+};
+
+struct tc_esfq_qopt
+{
+ unsigned quantum; /* Bytes per round allocated to flow */
+ int perturb_period; /* Period of hash perturbation */
+ __u32 limit; /* Maximal packets in queue */
+ unsigned divisor; /* Hash divisor */
+ unsigned flows; /* Maximal number of flows */
+ unsigned hash_kind; /* Hash function to use for flow identification */
+};
+
+
/* RED section */
enum
diff -urN --exclude=.svn iproute2-2.6.11-050330/tc/Makefile iproute2-2.6.11-050330/tc/Makefile
--- iproute2-2.6.11-050330/tc/Makefile 2007-05-04 22:21:48.000000000 -0400
+++ iproute2-2.6.11-050330/tc/Makefile 2007-05-04 22:27:37.000000000 -0400
@@ -6,6 +6,7 @@
diff -Naur iproute-2.6.20-070313.orig/tc/Makefile iproute-2.6.20-070313/tc/Makefile
--- iproute-2.6.20-070313.orig/tc/Makefile 2007-03-13 14:50:56.000000000 -0700
+++ iproute-2.6.20-070313/tc/Makefile 2007-06-09 00:39:44.000000000 -0700
@@ -7,6 +7,7 @@
TCMODULES :=
TCMODULES += q_fifo.o
TCMODULES += q_sfq.o
@ -89,10 +50,10 @@ diff -urN --exclude=.svn iproute2-2.6.11-050330/tc/Makefile iproute2-2.6.11-0503
TCMODULES += q_red.o
TCMODULES += q_prio.o
TCMODULES += q_tbf.o
diff -urN --exclude=.svn iproute2-2.6.11-050330/tc/q_esfq.c iproute2-2.6.11-050330/tc/q_esfq.c
--- iproute2-2.6.11-050330/tc/q_esfq.c 1969-12-31 19:00:00.000000000 -0500
+++ iproute2-2.6.11-050330/tc/q_esfq.c 2007-05-04 22:37:54.000000000 -0400
@@ -0,0 +1,200 @@
diff -Naur iproute-2.6.20-070313.orig/tc/q_esfq.c iproute-2.6.20-070313/tc/q_esfq.c
--- iproute-2.6.20-070313.orig/tc/q_esfq.c 1969-12-31 16:00:00.000000000 -0800
+++ iproute-2.6.20-070313/tc/q_esfq.c 2007-06-09 11:38:59.000000000 -0700
@@ -0,0 +1,198 @@
+/*
+ * q_esfq.c ESFQ.
+ *
@ -127,7 +88,7 @@ diff -urN --exclude=.svn iproute2-2.6.11-050330/tc/q_esfq.c iproute2-2.6.11-0503
+{
+ fprintf(stderr, "Usage: ... esfq [ perturb SECS ] [ quantum BYTES ] [ depth FLOWS ]\n\t[ divisor HASHBITS ] [ limit PKTS ] [ hash HASHTYPE]\n");
+ fprintf(stderr,"Where: \n");
+ fprintf(stderr,"HASHTYPE := { classic | src | dst | ctorigdst | ctorigsrc | ctrepldst | ctreplsrc | ctnatchg }\n");
+ fprintf(stderr,"HASHTYPE := { classic | src | dst | fwmark | ctorigdst | ctorigsrc | ctrepldst | ctreplsrc | ctnatchg}\n");
+}
+
+#define usage() return(-1)
@ -169,8 +130,8 @@ diff -urN --exclude=.svn iproute2-2.6.11-050330/tc/q_esfq.c iproute2-2.6.11-0503
+ fprintf(stderr, "Illegal \"divisor\"\n");
+ return -1;
+ }
+ if(opt.divisor >= 14) {
+ fprintf(stderr, "Illegal \"divisor\": must be < 14\n");
+ if(opt.divisor >= 15) {
+ fprintf(stderr, "Illegal \"divisor\": must be < 15\n");
+ return -1;
+ }
+ opt.divisor=pow(2,opt.divisor);
@ -186,26 +147,21 @@ diff -urN --exclude=.svn iproute2-2.6.11-050330/tc/q_esfq.c iproute2-2.6.11-0503
+ NEXT_ARG();
+ if (strcmp(*argv, "classic") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_CLASSIC;
+ } else
+ if(strcmp(*argv, "dst") == 0) {
+ } else if (strcmp(*argv, "dst") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_DST;
+ } else
+ if(strcmp(*argv, "src") == 0) {
+ } else if (strcmp(*argv, "src") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_SRC;
+ } else
+ if(strcmp(*argv, "ctorigsrc") == 0) {
+ } else if (strcmp(*argv, "fwmark") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_FWMARK;
+ } else if (strcmp(*argv, "ctorigsrc") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_CTORIGSRC;
+ } else
+ if(strcmp(*argv, "ctorigdst") == 0) {
+ } else if (strcmp(*argv, "ctorigdst") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_CTORIGDST;
+ } else
+ if(strcmp(*argv, "ctreplsrc") == 0) {
+ } else if (strcmp(*argv, "ctreplsrc") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_CTREPLSRC;
+ } else
+ if(strcmp(*argv, "ctrepldst") == 0) {
+ } else if (strcmp(*argv, "ctrepldst") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_CTREPLDST;
+ } else
+ if(strcmp(*argv, "ctnatchg") == 0) {
+ } else if (strcmp(*argv, "ctnatchg") == 0) {
+ opt.hash_kind = TCA_SFQ_HASH_CTNATCHG;
+ } else {
+ fprintf(stderr, "Illegal \"hash\"\n");
@ -260,6 +216,9 @@ diff -urN --exclude=.svn iproute2-2.6.11-050330/tc/q_esfq.c iproute2-2.6.11-0503
+ case TCA_SFQ_HASH_SRC:
+ fprintf(f,"src");
+ break;
+ case TCA_SFQ_HASH_FWMARK:
+ fprintf(f,"fwmark");
+ break;
+ case TCA_SFQ_HASH_CTORIGSRC:
+ fprintf(f,"ctorigsrc");
+ break;

View file

@ -1,13 +1,12 @@
Index: linux-2.6.23-rc6/include/linux/pkt_sched.h
===================================================================
--- linux-2.6.23-rc6.orig/include/linux/pkt_sched.h 2007-09-21 16:23:53.000000000 +0800
+++ linux-2.6.23-rc6/include/linux/pkt_sched.h 2007-09-21 16:24:04.000000000 +0800
@@ -155,8 +155,40 @@
diff -Naur linux-2.6.21.5.orig/include/linux/pkt_sched.h linux-2.6.21.5/include/linux/pkt_sched.h
--- linux-2.6.21.5.orig/include/linux/pkt_sched.h 2007-06-11 11:37:06.000000000 -0700
+++ linux-2.6.21.5/include/linux/pkt_sched.h 2007-06-22 22:53:46.000000000 -0700
@@ -146,8 +146,37 @@
*
* The only reason for this is efficiency, it is possible
* to change these parameters in compile time.
+ *
+ * If you need to play with these values use esfq instead.
+ * If you need to play with these values, use esfq instead.
*/
+/* ESFQ section */
@ -19,15 +18,12 @@ Index: linux-2.6.23-rc6/include/linux/pkt_sched.h
+ TCA_SFQ_HASH_DST,
+ TCA_SFQ_HASH_SRC,
+ TCA_SFQ_HASH_FWMARK,
+ /* direct */
+ TCA_SFQ_HASH_DSTDIR,
+ TCA_SFQ_HASH_SRCDIR,
+ TCA_SFQ_HASH_FWMARKDIR,
+ /* conntrack */
+ TCA_SFQ_HASH_CTORIGDST,
+ TCA_SFQ_HASH_CTORIGSRC,
+ TCA_SFQ_HASH_CTREPLDST,
+ TCA_SFQ_HASH_CTREPLSRC,
+ TCA_SFQ_HASH_CTNATCHG,
+};
+
+struct tc_esfq_qopt
@ -43,11 +39,10 @@ Index: linux-2.6.23-rc6/include/linux/pkt_sched.h
/* RED section */
enum
Index: linux-2.6.23-rc6/net/sched/Kconfig
===================================================================
--- linux-2.6.23-rc6.orig/net/sched/Kconfig 2007-09-21 16:23:53.000000000 +0800
+++ linux-2.6.23-rc6/net/sched/Kconfig 2007-09-21 16:24:04.000000000 +0800
@@ -144,6 +144,26 @@
diff -Naur linux-2.6.21.5.orig/net/sched/Kconfig linux-2.6.21.5/net/sched/Kconfig
--- linux-2.6.21.5.orig/net/sched/Kconfig 2007-06-11 11:37:06.000000000 -0700
+++ linux-2.6.21.5/net/sched/Kconfig 2007-06-23 14:11:02.000000000 -0700
@@ -189,6 +189,37 @@
To compile this code as a module, choose M here: the
module will be called sch_sfq.
@ -70,15 +65,25 @@ Index: linux-2.6.23-rc6/net/sched/Kconfig
+
+ To compile this code as a module, choose M here: the
+ module will be called sch_esfq.
+
+config NET_SCH_ESFQ_NFCT
+ bool "Connection Tracking Hash Types"
+ depends on NET_SCH_ESFQ && NF_CONNTRACK
+ ---help---
+ Say Y here to enable support for hashing based on netfilter connection
+ tracking information. This is useful for a router that is also using
+ NAT to connect privately-addressed hosts to the Internet. If you want
+ to provide fair distribution of upstream bandwidth, ESFQ must use
+ connection tracking information, since all outgoing packets will share
+ the same source address.
+
config NET_SCH_TEQL
tristate "True Link Equalizer (TEQL)"
---help---
Index: linux-2.6.23-rc6/net/sched/Makefile
===================================================================
--- linux-2.6.23-rc6.orig/net/sched/Makefile 2007-09-21 16:23:53.000000000 +0800
+++ linux-2.6.23-rc6/net/sched/Makefile 2007-09-21 16:24:04.000000000 +0800
@@ -22,6 +22,7 @@
diff -Naur linux-2.6.21.5.orig/net/sched/Makefile linux-2.6.21.5/net/sched/Makefile
--- linux-2.6.21.5.orig/net/sched/Makefile 2007-06-11 11:37:06.000000000 -0700
+++ linux-2.6.21.5/net/sched/Makefile 2007-06-22 22:53:46.000000000 -0700
@@ -23,6 +23,7 @@
obj-$(CONFIG_NET_SCH_INGRESS) += sch_ingress.o
obj-$(CONFIG_NET_SCH_DSMARK) += sch_dsmark.o
obj-$(CONFIG_NET_SCH_SFQ) += sch_sfq.o
@ -86,11 +91,10 @@ Index: linux-2.6.23-rc6/net/sched/Makefile
obj-$(CONFIG_NET_SCH_TBF) += sch_tbf.o
obj-$(CONFIG_NET_SCH_TEQL) += sch_teql.o
obj-$(CONFIG_NET_SCH_PRIO) += sch_prio.o
Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
===================================================================
--- /dev/null 1970-01-01 00:00:00.000000000 +0000
+++ linux-2.6.23-rc6/net/sched/sch_esfq.c 2007-09-21 16:24:04.000000000 +0800
@@ -0,0 +1,704 @@
diff -Naur linux-2.6.21.5.orig/net/sched/sch_esfq.c linux-2.6.21.5/net/sched/sch_esfq.c
--- linux-2.6.21.5.orig/net/sched/sch_esfq.c 1969-12-31 16:00:00.000000000 -0800
+++ linux-2.6.21.5/net/sched/sch_esfq.c 2007-06-23 19:18:00.000000000 -0700
@@ -0,0 +1,702 @@
+/*
+ * net/sched/sch_esfq.c Extended Stochastic Fairness Queueing discipline.
+ *
@ -111,9 +115,9 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ * Corey Hickey, <bugfood-c@fatooh.org>
+ * Maintenance of the Linux 2.6 port.
+ * Added fwmark hash (thanks to Robert Kurjata).
+ * Added direct hashing for src, dst, and fwmark.
+ * Added usage of jhash.
+ *
+ * Added conntrack support.
+ * Added ctnatchg hash (thanks to Ben Pfountz).
+ */
+
+#include <linux/module.h>
@ -143,10 +147,7 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+#include <net/sock.h>
+#include <net/pkt_sched.h>
+#include <linux/jhash.h>
+
+#ifdef CONFIG_NF_CONNTRACK_ENABLED
+#include <net/netfilter/nf_conntrack.h>
+#endif
+
+/* Stochastic Fairness Queuing algorithm.
+ For more comments look at sch_sfq.c.
@ -157,9 +158,6 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ dst: destination IP address
+ src: source IP address
+ fwmark: netfilter mark value
+ dst_direct:
+ src_direct:
+ fwmark_direct: direct hashing of the above sources
+ ctorigdst: original destination IP address
+ ctorigsrc: original source IP address
+ ctrepldst: reply destination IP address
@ -167,6 +165,8 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+
+*/
+
+#define ESFQ_HEAD 0
+#define ESFQ_TAIL 1
+
+/* This type should contain at least SFQ_DEPTH*2 values */
+typedef unsigned int esfq_index;
@ -198,9 +198,6 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ unsigned short *hash; /* Hash value indexed by slots */
+ struct sk_buff_head *qs; /* Slot queue */
+ struct esfq_head *dep; /* Linked list of slots, indexed by depth */
+ unsigned dyn_min; /* For dynamic divisor adjustment; minimum value seen */
+ unsigned dyn_max; /* maximum value seen */
+ unsigned dyn_range; /* saved range */
+};
+
+/* This contains the info we will hash. */
@ -216,31 +213,6 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ u32 mark; /* netfilter mark (fwmark) */
+};
+
+/* Hash input values directly into the "nearest" slot, taking into account the
+ * range of input values seen. This is most useful when the hash table is at
+ * least as large as the range of possible values.
+ * Note: this functionality was added before the change to using jhash, and may
+ * no longer be useful. */
+static __inline__ unsigned esfq_hash_direct(struct esfq_sched_data *q, u32 h)
+{
+ /* adjust minimum and maximum */
+ if (h < q->dyn_min || h > q->dyn_max) {
+ q->dyn_min = h < q->dyn_min ? h : q->dyn_min;
+ q->dyn_max = h > q->dyn_max ? h : q->dyn_max;
+
+ /* find new range */
+ if ((q->dyn_range = q->dyn_max - q->dyn_min) >= q->hash_divisor)
+ printk(KERN_WARNING "ESFQ: (direct hash) Input range %u is larger than hash "
+ "table. See ESFQ README for details.\n", q->dyn_range);
+ }
+
+ /* hash input values into slot numbers */
+ if (q->dyn_min == q->dyn_max)
+ return 0; /* only one value seen; avoid division by 0 */
+ else
+ return (h - q->dyn_min) * (q->hash_divisor - 1) / q->dyn_range;
+}
+
+static __inline__ unsigned esfq_jhash_1word(struct esfq_sched_data *q,u32 a)
+{
+ return jhash_1word(a, q->perturbation) & (q->hash_divisor-1);
@ -256,11 +228,10 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ return jhash_3words(a, b, c, q->perturbation) & (q->hash_divisor-1);
+}
+
+
+static unsigned esfq_hash(struct esfq_sched_data *q, struct sk_buff *skb)
+{
+ struct esfq_packet_info info;
+#ifdef CONFIG_NF_CONNTRACK_ENABLED
+#ifdef CONFIG_NET_SCH_ESFQ_NFCT
+ enum ip_conntrack_info ctinfo;
+ struct nf_conn *ct = nf_ct_get(skb, &ctinfo);
+#endif
@ -268,7 +239,7 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ switch (skb->protocol) {
+ case __constant_htons(ETH_P_IP):
+ {
+ struct iphdr *iph = ip_hdr(skb);
+ struct iphdr *iph = skb->nh.iph;
+ info.dst = iph->daddr;
+ info.src = iph->saddr;
+ if (!(iph->frag_off&htons(IP_MF|IP_OFFSET)) &&
@ -284,7 +255,7 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ }
+ case __constant_htons(ETH_P_IPV6):
+ {
+ struct ipv6hdr *iph = ipv6_hdr(skb);
+ struct ipv6hdr *iph = skb->nh.ipv6h;
+ /* Hash ipv6 addresses into a u32. This isn't ideal,
+ * but the code is simple. */
+ info.dst = jhash2(iph->daddr.s6_addr32, 4, q->perturbation);
@ -307,7 +278,7 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+
+ info.mark = skb->mark;
+
+#ifdef CONFIG_NF_CONNTRACK_ENABLED
+#ifdef CONFIG_NET_SCH_ESFQ_NFCT
+ /* defaults if there is no conntrack info */
+ info.ctorigsrc = info.src;
+ info.ctorigdst = info.dst;
@ -332,23 +303,16 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ }
+#endif
+
+ switch(q->hash_kind)
+ {
+ switch(q->hash_kind) {
+ case TCA_SFQ_HASH_CLASSIC:
+ return esfq_jhash_3words(q, info.dst, info.src, info.proto);
+ case TCA_SFQ_HASH_DST:
+ return esfq_jhash_1word(q, info.dst);
+ case TCA_SFQ_HASH_DSTDIR:
+ return esfq_hash_direct(q, ntohl(info.dst));
+ case TCA_SFQ_HASH_SRC:
+ return esfq_jhash_1word(q, info.src);
+ case TCA_SFQ_HASH_SRCDIR:
+ return esfq_hash_direct(q, ntohl(info.src));
+ case TCA_SFQ_HASH_FWMARK:
+ return esfq_jhash_1word(q, info.mark);
+ case TCA_SFQ_HASH_FWMARKDIR:
+ return esfq_hash_direct(q, info.mark);
+#ifdef CONFIG_NF_CONNTRACK_ENABLED
+#ifdef CONFIG_NET_SCH_ESFQ_NFCT
+ case TCA_SFQ_HASH_CTORIGDST:
+ return esfq_jhash_1word(q, info.ctorigdst);
+ case TCA_SFQ_HASH_CTORIGSRC:
@ -357,6 +321,12 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ return esfq_jhash_1word(q, info.ctrepldst);
+ case TCA_SFQ_HASH_CTREPLSRC:
+ return esfq_jhash_1word(q, info.ctreplsrc);
+ case TCA_SFQ_HASH_CTNATCHG:
+ {
+ if (info.ctorigdst == info.ctreplsrc)
+ return esfq_jhash_1word(q, info.ctorigsrc);
+ return esfq_jhash_1word(q, info.ctreplsrc);
+ }
+#endif
+ default:
+ if (net_ratelimit())
@ -451,10 +421,8 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ return 0;
+}
+
+static int
+esfq_enqueue(struct sk_buff *skb, struct Qdisc* sch)
+static void esfq_q_enqueue(struct sk_buff *skb, struct esfq_sched_data *q, unsigned int end)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ unsigned hash = esfq_hash(q, skb);
+ unsigned depth = q->depth;
+ esfq_index x;
@ -464,8 +432,12 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ q->ht[hash] = x = q->dep[depth].next;
+ q->hash[x] = hash;
+ }
+ sch->qstats.backlog += skb->len;
+
+ if (end == ESFQ_TAIL)
+ __skb_queue_tail(&q->qs[x], skb);
+ else
+ __skb_queue_head(&q->qs[x], skb);
+
+ esfq_inc(q, x);
+ if (q->qs[x].qlen == 1) { /* The flow is new */
+ if (q->tail == depth) { /* It is the first flow */
@ -478,43 +450,30 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ q->tail = x;
+ }
+ }
+}
+
+static int esfq_enqueue(struct sk_buff *skb, struct Qdisc* sch)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ esfq_q_enqueue(skb, q, ESFQ_TAIL);
+ sch->qstats.backlog += skb->len;
+ if (++sch->q.qlen < q->limit-1) {
+ sch->bstats.bytes += skb->len;
+ sch->bstats.packets++;
+ return 0;
+ }
+
+ sch->qstats.drops++;
+ esfq_drop(sch);
+ return NET_XMIT_CN;
+}
+
+static int
+esfq_requeue(struct sk_buff *skb, struct Qdisc* sch)
+
+static int esfq_requeue(struct sk_buff *skb, struct Qdisc* sch)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ unsigned hash = esfq_hash(q, skb);
+ unsigned depth = q->depth;
+ esfq_index x;
+
+ x = q->ht[hash];
+ if (x == depth) {
+ q->ht[hash] = x = q->dep[depth].next;
+ q->hash[x] = hash;
+ }
+ esfq_q_enqueue(skb, q, ESFQ_HEAD);
+ sch->qstats.backlog += skb->len;
+ __skb_queue_head(&q->qs[x], skb);
+ esfq_inc(q, x);
+ if (q->qs[x].qlen == 1) { /* The flow is new */
+ if (q->tail == 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;
@ -525,13 +484,8 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ return NET_XMIT_CN;
+}
+
+
+
+
+static struct sk_buff *
+esfq_dequeue(struct Qdisc* sch)
+static struct sk_buff *esfq_q_dequeue(struct esfq_sched_data *q)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ struct sk_buff *skb;
+ unsigned depth = q->depth;
+ esfq_index a, old_a;
@ -545,8 +499,6 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ /* Grab packet */
+ skb = __skb_dequeue(&q->qs[a]);
+ esfq_dec(q, a);
+ sch->q.qlen--;
+ sch->qstats.backlog -= skb->len;
+
+ /* Is the slot empty? */
+ if (q->qs[a].qlen == 0) {
@ -567,8 +519,44 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ return skb;
+}
+
+static void
+esfq_reset(struct Qdisc* sch)
+static struct sk_buff *esfq_dequeue(struct Qdisc* sch)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ struct sk_buff *skb;
+
+ skb = esfq_q_dequeue(q);
+ if (skb == NULL)
+ return NULL;
+ sch->q.qlen--;
+ sch->qstats.backlog -= skb->len;
+ return skb;
+}
+
+static void esfq_q_destroy(struct esfq_sched_data *q)
+{
+ del_timer(&q->perturb_timer);
+ if(q->ht)
+ kfree(q->ht);
+ if(q->dep)
+ kfree(q->dep);
+ if(q->next)
+ kfree(q->next);
+ if(q->allot)
+ kfree(q->allot);
+ if(q->hash)
+ kfree(q->hash);
+ if(q->qs)
+ kfree(q->qs);
+}
+
+static void esfq_destroy(struct Qdisc *sch)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ esfq_q_destroy(q);
+}
+
+
+static void esfq_reset(struct Qdisc* sch)
+{
+ struct sk_buff *skb;
+
@ -589,73 +577,56 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ }
+}
+
+static int esfq_change(struct Qdisc *sch, struct rtattr *opt)
+static unsigned int esfq_check_hash(unsigned int kind)
+{
+ switch (kind) {
+ case TCA_SFQ_HASH_CTORIGDST:
+ case TCA_SFQ_HASH_CTORIGSRC:
+ case TCA_SFQ_HASH_CTREPLDST:
+ case TCA_SFQ_HASH_CTREPLSRC:
+ case TCA_SFQ_HASH_CTNATCHG:
+#ifndef CONFIG_NET_SCH_ESFQ_NFCT
+ {
+ if (net_ratelimit())
+ printk(KERN_WARNING "ESFQ: Conntrack hash types disabled in kernel config. Falling back to classic.\n");
+ return TCA_SFQ_HASH_CLASSIC;
+ }
+#endif
+ case TCA_SFQ_HASH_CLASSIC:
+ case TCA_SFQ_HASH_DST:
+ case TCA_SFQ_HASH_SRC:
+ case TCA_SFQ_HASH_FWMARK:
+ return kind;
+ default:
+ {
+ if (net_ratelimit())
+ printk(KERN_WARNING "ESFQ: Unknown hash type. Falling back to classic.\n");
+ return TCA_SFQ_HASH_CLASSIC;
+ }
+ }
+}
+
+static int esfq_q_init(struct esfq_sched_data *q, struct rtattr *opt)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ struct tc_esfq_qopt *ctl = RTA_DATA(opt);
+ int old_perturb = q->perturb_period;
+
+ 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;
+// q->hash_divisor = ctl->divisor;
+// q->tail = q->limit = q->depth = ctl->flows;
+
+ if (ctl->limit)
+ q->limit = min_t(u32, ctl->limit, q->depth);
+
+ if (ctl->hash_kind) {
+ q->hash_kind = ctl->hash_kind;
+ if (q->hash_kind != TCA_SFQ_HASH_CLASSIC)
+ q->perturb_period = 0;
+ }
+
+ // is sch_tree_lock enough to do this ?
+ while (sch->q.qlen >= q->limit-1)
+ esfq_drop(sch);
+
+ if (old_perturb)
+ del_timer(&q->perturb_timer);
+ if (q->perturb_period) {
+ q->perturb_timer.expires = jiffies + q->perturb_period;
+ add_timer(&q->perturb_timer);
+ } else {
+ q->perturbation = 0;
+ }
+ sch_tree_unlock(sch);
+ return 0;
+}
+
+static int esfq_init(struct Qdisc *sch, struct rtattr *opt)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ struct tc_esfq_qopt *ctl;
+ esfq_index p = ~0U/2;
+ int i;
+
+ if (opt && opt->rta_len < RTA_LENGTH(sizeof(*ctl)))
+ return -EINVAL;
+
+ init_timer(&q->perturb_timer);
+ q->perturb_timer.data = (unsigned long)sch;
+ q->perturb_timer.function = esfq_perturbation;
+ q->perturbation = 0;
+ q->hash_kind = TCA_SFQ_HASH_CLASSIC;
+ q->max_depth = 0;
+ q->dyn_min = ~0U; /* maximum value for this type */
+ q->dyn_max = 0; /* dyn_min/dyn_max will be set properly upon first packet */
+ if (opt == NULL) {
+ q->quantum = psched_mtu(sch->dev);
+ q->perturb_period = 0;
+ q->hash_divisor = 1024;
+ q->tail = q->limit = q->depth = 128;
+
+ } else {
+ ctl = RTA_DATA(opt);
+ q->quantum = ctl->quantum ? : psched_mtu(sch->dev);
+ struct tc_esfq_qopt *ctl = RTA_DATA(opt);
+ if (ctl->quantum)
+ q->quantum = ctl->quantum;
+ q->perturb_period = ctl->perturb_period*HZ;
+ q->hash_divisor = ctl->divisor ? : 1024;
+ q->tail = q->limit = q->depth = ctl->flows ? : 128;
@ -667,26 +638,19 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ q->limit = min_t(u32, ctl->limit, q->depth);
+
+ if (ctl->hash_kind) {
+ q->hash_kind = ctl->hash_kind;
+ }
+
+ if (q->perturb_period) {
+ q->perturb_timer.expires = jiffies + q->perturb_period;
+ add_timer(&q->perturb_timer);
+ q->hash_kind = esfq_check_hash(ctl->hash_kind);
+ }
+ }
+
+ q->ht = kmalloc(q->hash_divisor*sizeof(esfq_index), GFP_KERNEL);
+ if (!q->ht)
+ goto err_case;
+
+ q->dep = kmalloc((1+q->depth*2)*sizeof(struct esfq_head), GFP_KERNEL);
+ if (!q->dep)
+ goto err_case;
+ q->next = kmalloc(q->depth*sizeof(esfq_index), GFP_KERNEL);
+ if (!q->next)
+ goto err_case;
+
+ q->allot = kmalloc(q->depth*sizeof(short), GFP_KERNEL);
+ if (!q->allot)
+ goto err_case;
@ -709,38 +673,76 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ esfq_link(q, i);
+ return 0;
+err_case:
+ del_timer(&q->perturb_timer);
+ if (q->ht)
+ kfree(q->ht);
+ if (q->dep)
+ kfree(q->dep);
+ if (q->next)
+ kfree(q->next);
+ if (q->allot)
+ kfree(q->allot);
+ if (q->hash)
+ kfree(q->hash);
+ if (q->qs)
+ kfree(q->qs);
+ esfq_q_destroy(q);
+ return -ENOBUFS;
+}
+
+static void esfq_destroy(struct Qdisc *sch)
+static int esfq_init(struct Qdisc *sch, struct rtattr *opt)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ del_timer(&q->perturb_timer);
+ if(q->ht)
+ kfree(q->ht);
+ if(q->dep)
+ kfree(q->dep);
+ if(q->next)
+ kfree(q->next);
+ if(q->allot)
+ kfree(q->allot);
+ if(q->hash)
+ kfree(q->hash);
+ if(q->qs)
+ kfree(q->qs);
+ int err;
+
+ q->quantum = psched_mtu(sch->dev); /* default */
+ if ((err = esfq_q_init(q, opt)))
+ return err;
+
+ init_timer(&q->perturb_timer);
+ q->perturb_timer.data = (unsigned long)sch;
+ q->perturb_timer.function = esfq_perturbation;
+ if (q->perturb_period) {
+ q->perturb_timer.expires = jiffies + q->perturb_period;
+ add_timer(&q->perturb_timer);
+ }
+
+ return 0;
+}
+
+static int esfq_change(struct Qdisc *sch, struct rtattr *opt)
+{
+ struct esfq_sched_data *q = qdisc_priv(sch);
+ struct esfq_sched_data new;
+ struct sk_buff *skb;
+ int err;
+
+ /* set up new queue */
+ memset(&new, 0, sizeof(struct esfq_sched_data));
+ new.quantum = psched_mtu(sch->dev); /* default */
+ if ((err = esfq_q_init(&new, opt)))
+ return err;
+
+ /* copy all packets from the old queue to the new queue */
+ sch_tree_lock(sch);
+ while ((skb = esfq_q_dequeue(q)) != NULL)
+ esfq_q_enqueue(skb, &new, ESFQ_TAIL);
+
+ /* clean up the old queue */
+ esfq_q_destroy(q);
+
+ /* copy elements of the new queue into the old queue */
+ q->perturb_period = new.perturb_period;
+ q->quantum = new.quantum;
+ q->limit = new.limit;
+ q->depth = new.depth;
+ q->hash_divisor = new.hash_divisor;
+ q->hash_kind = new.hash_kind;
+ q->tail = new.tail;
+ q->max_depth = new.max_depth;
+ q->ht = new.ht;
+ q->dep = new.dep;
+ q->next = new.next;
+ q->allot = new.allot;
+ q->hash = new.hash;
+ q->qs = new.qs;
+
+ /* finish up */
+ if (q->perturb_period) {
+ q->perturb_timer.expires = jiffies + q->perturb_period;
+ add_timer(&q->perturb_timer);
+ } else {
+ q->perturbation = 0;
+ }
+ sch_tree_unlock(sch);
+ return 0;
+}
+
+static int esfq_dump(struct Qdisc *sch, struct sk_buff *skb)
@ -779,7 +781,7 @@ Index: linux-2.6.23-rc6/net/sched/sch_esfq.c
+ .init = esfq_init,
+ .reset = esfq_reset,
+ .destroy = esfq_destroy,
+ .change = NULL, /* esfq_change - needs more work */
+ .change = esfq_change,
+ .dump = esfq_dump,
+ .owner = THIS_MODULE,
+};