IPFire-SCM August 2012

ipfire-scm@lists.ipfire.org

1 participants
50 discussions

[git.ipfire.org] IPFire 2.x development tree branch, kernel-update, updated. 38bb1dd6653f79102e17aaae7e72ce85dbc589d6
by git＠ipfire.org 20 Aug '12

20 Aug '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "IPFire 2.x development tree". The branch, kernel-update has been updated via 38bb1dd6653f79102e17aaae7e72ce85dbc589d6 (commit) via 73a1c982baf987ecc1b8576dba473f79bd0ff926 (commit) from 3e78d712ccafa64c78439c02d2fae5a3615b0cd5 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 38bb1dd6653f79102e17aaae7e72ce85dbc589d6 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Sat Aug 18 19:22:21 2012 +0200 compat-wireless: update to 3.5.1-1-snpc. commit 73a1c982baf987ecc1b8576dba473f79bd0ff926 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Sat Aug 18 17:13:53 2012 +0200 Revert "kernel: add backported codel patchset." This reverts commit 97bb1c57aea0cfdb554a7bb96d1a121a868b6fa3. compat-wireless provide the same functionality. ----------------------------------------------------------------------- Summary of changes: .../kernel/kernel.config.armv5tel-ipfire-kirkwood | 2 - config/kernel/kernel.config.armv5tel-ipfire-omap | 2 - config/kernel/kernel.config.armv5tel-ipfire-rpi | 2 - config/kernel/kernel.config.i586-ipfire | 4 +- config/kernel/kernel.config.i586-ipfire-pae | 4 +- lfs/compat-wireless | 7 +- lfs/linux | 3 - ...t-wireless_codel-avoid-a-nul-rec_inv_sqrt.patch | 68 ++ .../0000-net-introduce-skb_flow_dissect.patch | 195 ----- .../0001-codel-Controlled-Delay-AQM.patch | 771 ------------------ ...Newton-method-instead-of-sqrt-and-divides.patch | 189 ----- .../0003-fq_codel-Fair-Queue-Codel-AQM.patch | 850 -------------------- ...odel-Add-missing-include-linux-prefetch.h.patch | 37 - .../0005-net-codel-fix-build-errors.patch | 55 -- ...16-field-instead-of-31bits-for-rec_inv_sq.patch | 90 -- ...del-should-use-qdisc-backlog-as-threshold.patch | 140 ---- ...008-flow_dissector-use-a-64bit-load-store.patch | 90 -- ...issector.c-missing-include-linux-export.h.patch | 25 - 18 files changed, 75 insertions(+), 2459 deletions(-) create mode 100644 src/patches/compat-wireless_codel-avoid-a-nul-rec_inv_sqrt.patch delete mode 100644 src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch delete mode 100644 src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch delete mode 100644 src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch delete mode 100644 src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch delete mode 100644 src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch delete mode 100644 src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch delete mode 100644 src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch delete mode 100644 src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch delete mode 100644 src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch delete mode 100644 src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch Difference in files: diff --git a/config/kernel/kernel.config.armv5tel-ipfire-kirkwood b/config/kernel/kernel.config.armv5tel-ipfire-kirkwood index 67992fd..133ac1a 100644 --- a/config/kernel/kernel.config.armv5tel-ipfire-kirkwood +++ b/config/kernel/kernel.config.armv5tel-ipfire-kirkwood @@ -832,8 +832,6 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m -CONFIG_NET_SCH_CODEL=m -CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.armv5tel-ipfire-omap b/config/kernel/kernel.config.armv5tel-ipfire-omap index 17d5109..08405e2 100644 --- a/config/kernel/kernel.config.armv5tel-ipfire-omap +++ b/config/kernel/kernel.config.armv5tel-ipfire-omap @@ -924,8 +924,6 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m -CONFIG_NET_SCH_CODEL=m -CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.armv5tel-ipfire-rpi b/config/kernel/kernel.config.armv5tel-ipfire-rpi index 414dd8f..4d434e0 100644 --- a/config/kernel/kernel.config.armv5tel-ipfire-rpi +++ b/config/kernel/kernel.config.armv5tel-ipfire-rpi @@ -798,8 +798,6 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m -CONFIG_NET_SCH_CODEL=m -CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.i586-ipfire b/config/kernel/kernel.config.i586-ipfire index 1581ea0..4295eb4 100644 --- a/config/kernel/kernel.config.i586-ipfire +++ b/config/kernel/kernel.config.i586-ipfire @@ -1,6 +1,6 @@ # # Automatically generated file; DO NOT EDIT. -# Linux/i386 3.2.27 Kernel Configuration +# Linux/i386 3.2.20 Kernel Configuration # # CONFIG_64BIT is not set CONFIG_X86_32=y @@ -973,8 +973,6 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m -CONFIG_NET_SCH_CODEL=m -CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.i586-ipfire-pae b/config/kernel/kernel.config.i586-ipfire-pae index 4157b09..7186e73 100644 --- a/config/kernel/kernel.config.i586-ipfire-pae +++ b/config/kernel/kernel.config.i586-ipfire-pae @@ -1,6 +1,6 @@ # # Automatically generated file; DO NOT EDIT. -# Linux/i386 3.2.27 Kernel Configuration +# Linux/i386 2.6.42.20 Kernel Configuration # # CONFIG_64BIT is not set CONFIG_X86_32=y @@ -996,8 +996,6 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m -CONFIG_NET_SCH_CODEL=m -CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/lfs/compat-wireless b/lfs/compat-wireless index 24ad75b..78bd1d4 100644 --- a/lfs/compat-wireless +++ b/lfs/compat-wireless @@ -26,7 +26,7 @@ include Config VERSUFIX=ipfire$(KCFG) -VER = 3.5-3-snpc +VER = 3.5.1-1-snpc ifeq "$(KCFG)" "-xen" KVER = 2.6.32.59 @@ -47,7 +47,7 @@ objects = $(DL_FILE) asix-4.4.0.tar.xz $(DL_FILE) = $(DL_FROM)/$(DL_FILE) asix-4.4.0.tar.xz = $(DL_FROM)/asix-4.4.0.tar.xz -$(DL_FILE)_MD5 = 66f27eed39aacd567f67025305273cd7 +$(DL_FILE)_MD5 = eea8d1c430ad38b713457438e04f47f5 asix-4.4.0.tar.xz_MD5=633609e889de41554826e0e2cd7bffde install : $(TARGET) @@ -82,6 +82,9 @@ $(TARGET) : $(patsubst %,$(DIR_DL)/%,$(objects)) # kfifo has no license info and taints kernel cd $(DIR_APP) && patch -Np1 < $(DIR_SRC)/src/patches/compat-wireless-2.6.39_kfifo_module_info.patch + # Codel patches + cd $(DIR_APP) && patch -Np1 < $(DIR_SRC)/src/patches/compat-wireless_codel-avoid-a-nul-rec_inv_sqrt.patch + # Build ath5k only if target has pci cd $(DIR_APP) && patch -Np1 < $(DIR_SRC)/src/patches/compat-wireless-3.5-build_ath5k_only_with_pci.patch diff --git a/lfs/linux b/lfs/linux index 2c92809..9f994f8 100644 --- a/lfs/linux +++ b/lfs/linux @@ -111,9 +111,6 @@ $(TARGET) : $(patsubst %,$(DIR_DL)/%,$(objects)) ln -svf linux-$(VER) $(DIR_SRC)/linux - # Apply backported codel patches - cd $(DIR_APP) && cat $(DIR_SRC)/src/patches/linux-3.2-codel/*.patch | patch -Np1 - # Linux Intermediate Queueing Device cd $(DIR_APP) && patch -Np1 < $(DIR_SRC)/src/patches/linux-3.2-imq.patch diff --git a/src/patches/compat-wireless_codel-avoid-a-nul-rec_inv_sqrt.patch b/src/patches/compat-wireless_codel-avoid-a-nul-rec_inv_sqrt.patch new file mode 100644 index 0000000..210a58c --- /dev/null +++ b/src/patches/compat-wireless_codel-avoid-a-nul-rec_inv_sqrt.patch @@ -0,0 +1,68 @@ +From patchwork Mon Jul 30 06:52:21 2012 +Content-Type: text/plain; charset="utf-8" +MIME-Version: 1.0 +Content-Transfer-Encoding: 7bit +Subject: codel: refine one condition to avoid a nul rec_inv_sqrt +Date: Sun, 29 Jul 2012 20:52:21 -0000 +From: Eric Dumazet <eric.dumazet(a)gmail.com> +X-Patchwork-Id: 173968 +Message-Id: <1343631141.2626.13293.camel@edumazet-glaptop> +To: David Miller <davem(a)davemloft.net> +Cc: netdev <netdev(a)vger.kernel.org>, Anton Mich <lp2s1h(a)gmail.com> + +From: Eric Dumazet <edumazet(a)google.com> + +One condition before codel_Newton_step() was not good if +we never left the dropping state for a flow. As a result +rec_inv_sqrt was 0, instead of the ~0 initial value. + +codel control law was then set to a very aggressive mode, dropping +many packets before reaching 'target' and recovering from this problem. + +To keep codel_vars_init() as efficient as possible, refine +the condition to make sure rec_inv_sqrt initial value is correct + +Many thanks to Anton Mich for discovering the issue and suggesting +a fix. + +Reported-by: Anton Mich <lp2s1h(a)gmail.com> +Signed-off-by: Eric Dumazet <edumazet(a)google.com> + +--- +include/net/codel.h | 8 ++++++-- + 1 file changed, 6 insertions(+), 2 deletions(-) + + + +-- +To unsubscribe from this list: send the line "unsubscribe netdev" in +the body of a message to majordomo(a)vger.kernel.org +More majordomo info at http://vger.kernel.org/majordomo-info.html + +diff --git a/include/net/codel.h b/include/net/codel.h +index 550debf..389cf62 100644 +--- a/include/net/codel.h ++++ b/include/net/codel.h +@@ -305,6 +305,8 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + } + } + } else if (drop) { ++ u32 delta; ++ + if (params->ecn && INET_ECN_set_ce(skb)) { + stats->ecn_mark++; + } else { +@@ -320,9 +322,11 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + * assume that the drop rate that controlled the queue on the + * last cycle is a good starting point to control it now. + */ +- if (codel_time_before(now - vars->drop_next, ++ delta = vars->count - vars->lastcount; ++ if (delta > 1 && ++ codel_time_before(now - vars->drop_next, + 16 * params->interval)) { +- vars->count = (vars->count - vars->lastcount) | 1; ++ vars->count = delta; + /* we dont care if rec_inv_sqrt approximation + * is not very precise : + * Next Newton steps will correct it quadratically. diff --git a/src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch b/src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch deleted file mode 100644 index a749507..0000000 --- a/src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch +++ /dev/null @@ -1,195 +0,0 @@ -From: Eric Dumazet <eric.dumazet(a)gmail.com> -Date: Mon, 28 Nov 2011 05:22:18 +0000 -Subject: [PATCH 1/3] net: introduce skb_flow_dissect() - -commit 0744dd00c1b1be99a25b62b1b48df440e82e57e0 upstream. - -We use at least two flow dissectors in network stack, with known -limitations and code duplication. - -Introduce skb_flow_dissect() to factorize this, highly inspired from -existing dissector from __skb_get_rxhash() - -Note : We extensively use skb_header_pointer(), this permits us to not -touch skb at all. - -Signed-off-by: Eric Dumazet <eric.dumazet(a)gmail.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> - -diff --git a/include/net/flow_keys.h b/include/net/flow_keys.h -new file mode 100644 -index 0000000..e4cb285 ---- /dev/null -+++ b/include/net/flow_keys.h -@@ -0,0 +1,15 @@ -+#ifndef _NET_FLOW_KEYS_H -+#define _NET_FLOW_KEYS_H -+ -+struct flow_keys { -+ __be32 src; -+ __be32 dst; -+ union { -+ __be32 ports; -+ __be16 port16[2]; -+ }; -+ u8 ip_proto; -+}; -+ -+extern bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow); -+#endif -diff --git a/net/core/Makefile b/net/core/Makefile -index 3606d40..c4ecc86 100644 ---- a/net/core/Makefile -+++ b/net/core/Makefile -@@ -3,7 +3,7 @@ - # - - obj-y := sock.o request_sock.o skbuff.o iovec.o datagram.o stream.o scm.o \ -- gen_stats.o gen_estimator.o net_namespace.o secure_seq.o -+ gen_stats.o gen_estimator.o net_namespace.o secure_seq.o flow_dissector.o - - obj-$(CONFIG_SYSCTL) += sysctl_net_core.o - -diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c -new file mode 100644 -index 0000000..f0516d9 ---- /dev/null -+++ b/net/core/flow_dissector.c -@@ -0,0 +1,134 @@ -+#include <linux/skbuff.h> -+#include <linux/ip.h> -+#include <linux/ipv6.h> -+#include <linux/if_vlan.h> -+#include <net/ip.h> -+#include <linux/if_tunnel.h> -+#include <linux/if_pppox.h> -+#include <linux/ppp_defs.h> -+#include <net/flow_keys.h> -+ -+ -+bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow) -+{ -+ int poff, nhoff = skb_network_offset(skb); -+ u8 ip_proto; -+ __be16 proto = skb->protocol; -+ -+ memset(flow, 0, sizeof(*flow)); -+ -+again: -+ switch (proto) { -+ case __constant_htons(ETH_P_IP): { -+ const struct iphdr *iph; -+ struct iphdr _iph; -+ip: -+ iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph); -+ if (!iph) -+ return false; -+ -+ if (ip_is_fragment(iph)) -+ ip_proto = 0; -+ else -+ ip_proto = iph->protocol; -+ flow->src = iph->saddr; -+ flow->dst = iph->daddr; -+ nhoff += iph->ihl * 4; -+ break; -+ } -+ case __constant_htons(ETH_P_IPV6): { -+ const struct ipv6hdr *iph; -+ struct ipv6hdr _iph; -+ipv6: -+ iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph); -+ if (!iph) -+ return false; -+ -+ ip_proto = iph->nexthdr; -+ flow->src = iph->saddr.s6_addr32[3]; -+ flow->dst = iph->daddr.s6_addr32[3]; -+ nhoff += sizeof(struct ipv6hdr); -+ break; -+ } -+ case __constant_htons(ETH_P_8021Q): { -+ const struct vlan_hdr *vlan; -+ struct vlan_hdr _vlan; -+ -+ vlan = skb_header_pointer(skb, nhoff, sizeof(_vlan), &_vlan); -+ if (!vlan) -+ return false; -+ -+ proto = vlan->h_vlan_encapsulated_proto; -+ nhoff += sizeof(*vlan); -+ goto again; -+ } -+ case __constant_htons(ETH_P_PPP_SES): { -+ struct { -+ struct pppoe_hdr hdr; -+ __be16 proto; -+ } *hdr, _hdr; -+ hdr = skb_header_pointer(skb, nhoff, sizeof(_hdr), &_hdr); -+ if (!hdr) -+ return false; -+ proto = hdr->proto; -+ nhoff += PPPOE_SES_HLEN; -+ switch (proto) { -+ case __constant_htons(PPP_IP): -+ goto ip; -+ case __constant_htons(PPP_IPV6): -+ goto ipv6; -+ default: -+ return false; -+ } -+ } -+ default: -+ return false; -+ } -+ -+ switch (ip_proto) { -+ case IPPROTO_GRE: { -+ struct gre_hdr { -+ __be16 flags; -+ __be16 proto; -+ } *hdr, _hdr; -+ -+ hdr = skb_header_pointer(skb, nhoff, sizeof(_hdr), &_hdr); -+ if (!hdr) -+ return false; -+ /* -+ * Only look inside GRE if version zero and no -+ * routing -+ */ -+ if (!(hdr->flags & (GRE_VERSION|GRE_ROUTING))) { -+ proto = hdr->proto; -+ nhoff += 4; -+ if (hdr->flags & GRE_CSUM) -+ nhoff += 4; -+ if (hdr->flags & GRE_KEY) -+ nhoff += 4; -+ if (hdr->flags & GRE_SEQ) -+ nhoff += 4; -+ goto again; -+ } -+ break; -+ } -+ case IPPROTO_IPIP: -+ goto again; -+ default: -+ break; -+ } -+ -+ flow->ip_proto = ip_proto; -+ poff = proto_ports_offset(ip_proto); -+ if (poff >= 0) { -+ __be32 *ports, _ports; -+ -+ nhoff += poff; -+ ports = skb_header_pointer(skb, nhoff, sizeof(_ports), &_ports); -+ if (ports) -+ flow->ports = *ports; -+ } -+ -+ return true; -+} -+EXPORT_SYMBOL(skb_flow_dissect); --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch b/src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch deleted file mode 100644 index 0d8d18b..0000000 --- a/src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch +++ /dev/null @@ -1,771 +0,0 @@ -From: Eric Dumazet <edumazet(a)google.com> -Date: Thu, 10 May 2012 07:51:25 +0000 -Subject: [PATCH 1/7] codel: Controlled Delay AQM - -commit 76e3cc126bb223013a6b9a0e2a51238d1ef2e409 upstream. - -An implementation of CoDel AQM, from Kathleen Nichols and Van Jacobson. - -http://queue.acm.org/detail.cfm?id=2209336 - -This AQM main input is no longer queue size in bytes or packets, but the -delay packets stay in (FIFO) queue. - -As we don't have infinite memory, we still can drop packets in enqueue() -in case of massive load, but mean of CoDel is to drop packets in -dequeue(), using a control law based on two simple parameters : - -target : target sojourn time (default 5ms) -interval : width of moving time window (default 100ms) - -Based on initial work from Dave Taht. - -Refactored to help future codel inclusion as a plugin for other linux -qdisc (FQ_CODEL, ...), like RED. - -include/net/codel.h contains codel algorithm as close as possible than -Kathleen reference. - -net/sched/sch_codel.c contains the linux qdisc specific glue. - -Separate structures permit a memory efficient implementation of fq_codel -(to be sent as a separate work) : Each flow has its own struct -codel_vars. - -timestamps are taken at enqueue() time with 1024 ns precision, allowing -a range of 2199 seconds in queue, and 100Gb links support. iproute2 uses -usec as base unit. - -Selected packets are dropped, unless ECN is enabled and packets can get -ECN mark instead. - -Tested from 2Mb to 10Gb speeds with no particular problems, on ixgbe and -tg3 drivers (BQL enabled). - -Usage: tc qdisc ... codel [ limit PACKETS ] [ target TIME ] - [ interval TIME ] [ ecn ] - -qdisc codel 10: parent 1:1 limit 2000p target 3.0ms interval 60.0ms ecn - Sent 13347099587 bytes 8815805 pkt (dropped 0, overlimits 0 requeues 0) - rate 202365Kbit 16708pps backlog 113550b 75p requeues 0 - count 116 lastcount 98 ldelay 4.3ms dropping drop_next 816us - maxpacket 1514 ecn_mark 84399 drop_overlimit 0 - -CoDel must be seen as a base module, and should be used keeping in mind -there is still a FIFO queue. So a typical setup will probably need a -hierarchy of several qdiscs and packet classifiers to be able to meet -whatever constraints a user might have. - -One possible example would be to use fq_codel, which combines Fair -Queueing and CoDel, in replacement of sfq / sfq_red. - -Signed-off-by: Eric Dumazet <edumazet(a)google.com> -Signed-off-by: Dave Taht <dave.taht(a)bufferbloat.net> -Cc: Kathleen Nichols <nichols(a)pollere.com> -Cc: Van Jacobson <van(a)pollere.net> -Cc: Tom Herbert <therbert(a)google.com> -Cc: Matt Mathis <mattmathis(a)google.com> -Cc: Yuchung Cheng <ycheng(a)google.com> -Cc: Stephen Hemminger <shemminger(a)vyatta.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> ---- - include/linux/pkt_sched.h | 26 ++++ - include/net/codel.h | 332 +++++++++++++++++++++++++++++++++++++++++++++ - net/sched/Kconfig | 11 ++ - net/sched/Makefile | 1 + - net/sched/sch_codel.c | 275 +++++++++++++++++++++++++++++++++++++ - 5 files changed, 645 insertions(+) - create mode 100644 include/net/codel.h - create mode 100644 net/sched/sch_codel.c - -diff --git a/include/linux/pkt_sched.h b/include/linux/pkt_sched.h -index ffe975c..cde56c2 100644 ---- a/include/linux/pkt_sched.h -+++ b/include/linux/pkt_sched.h -@@ -655,4 +655,30 @@ struct tc_qfq_stats { - __u32 lmax; - }; - -+/* CODEL */ -+ -+enum { -+ TCA_CODEL_UNSPEC, -+ TCA_CODEL_TARGET, -+ TCA_CODEL_LIMIT, -+ TCA_CODEL_INTERVAL, -+ TCA_CODEL_ECN, -+ __TCA_CODEL_MAX -+}; -+ -+#define TCA_CODEL_MAX (__TCA_CODEL_MAX - 1) -+ -+struct tc_codel_xstats { -+ __u32 maxpacket; /* largest packet we've seen so far */ -+ __u32 count; /* how many drops we've done since the last time we -+ * entered dropping state -+ */ -+ __u32 lastcount; /* count at entry to dropping state */ -+ __u32 ldelay; /* in-queue delay seen by most recently dequeued packet */ -+ __s32 drop_next; /* time to drop next packet */ -+ __u32 drop_overlimit; /* number of time max qdisc packet limit was hit */ -+ __u32 ecn_mark; /* number of packets we ECN marked instead of dropped */ -+ __u32 dropping; /* are we in dropping state ? */ -+}; -+ - #endif -diff --git a/include/net/codel.h b/include/net/codel.h -new file mode 100644 -index 0000000..bce2cef ---- /dev/null -+++ b/include/net/codel.h -@@ -0,0 +1,332 @@ -+#ifndef __NET_SCHED_CODEL_H -+#define __NET_SCHED_CODEL_H -+ -+/* -+ * Codel - The Controlled-Delay Active Queue Management algorithm -+ * -+ * Copyright (C) 2011-2012 Kathleen Nichols <nichols(a)pollere.com> -+ * Copyright (C) 2011-2012 Van Jacobson <van(a)pollere.net> -+ * Copyright (C) 2012 Michael D. Taht <dave.taht(a)bufferbloat.net> -+ * Copyright (C) 2012 Eric Dumazet <edumazet(a)google.com> -+ * -+ * Redistribution and use in source and binary forms, with or without -+ * modification, are permitted provided that the following conditions -+ * are met: -+ * 1. Redistributions of source code must retain the above copyright -+ * notice, this list of conditions, and the following disclaimer, -+ * without modification. -+ * 2. Redistributions in binary form must reproduce the above copyright -+ * notice, this list of conditions and the following disclaimer in the -+ * documentation and/or other materials provided with the distribution. -+ * 3. The names of the authors may not be used to endorse or promote products -+ * derived from this software without specific prior written permission. -+ * -+ * Alternatively, provided that this notice is retained in full, this -+ * software may be distributed under the terms of the GNU General -+ * Public License ("GPL") version 2, in which case the provisions of the -+ * GPL apply INSTEAD OF those given above. -+ * -+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH -+ * DAMAGE. -+ * -+ */ -+ -+#include <linux/types.h> -+#include <linux/ktime.h> -+#include <linux/skbuff.h> -+#include <net/pkt_sched.h> -+#include <net/inet_ecn.h> -+ -+/* Controlling Queue Delay (CoDel) algorithm -+ * ========================================= -+ * Source : Kathleen Nichols and Van Jacobson -+ * http://queue.acm.org/detail.cfm?id=2209336 -+ * -+ * Implemented on linux by Dave Taht and Eric Dumazet -+ */ -+ -+ -+/* CoDel uses a 1024 nsec clock, encoded in u32 -+ * This gives a range of 2199 seconds, because of signed compares -+ */ -+typedef u32 codel_time_t; -+typedef s32 codel_tdiff_t; -+#define CODEL_SHIFT 10 -+#define MS2TIME(a) ((a * NSEC_PER_MSEC) >> CODEL_SHIFT) -+ -+static inline codel_time_t codel_get_time(void) -+{ -+ u64 ns = ktime_to_ns(ktime_get()); -+ -+ return ns >> CODEL_SHIFT; -+} -+ -+#define codel_time_after(a, b) ((s32)(a) - (s32)(b) > 0) -+#define codel_time_after_eq(a, b) ((s32)(a) - (s32)(b) >= 0) -+#define codel_time_before(a, b) ((s32)(a) - (s32)(b) < 0) -+#define codel_time_before_eq(a, b) ((s32)(a) - (s32)(b) <= 0) -+ -+/* Qdiscs using codel plugin must use codel_skb_cb in their own cb[] */ -+struct codel_skb_cb { -+ codel_time_t enqueue_time; -+}; -+ -+static struct codel_skb_cb *get_codel_cb(const struct sk_buff *skb) -+{ -+ qdisc_cb_private_validate(skb, sizeof(struct codel_skb_cb)); -+ return (struct codel_skb_cb *)qdisc_skb_cb(skb)->data; -+} -+ -+static codel_time_t codel_get_enqueue_time(const struct sk_buff *skb) -+{ -+ return get_codel_cb(skb)->enqueue_time; -+} -+ -+static void codel_set_enqueue_time(struct sk_buff *skb) -+{ -+ get_codel_cb(skb)->enqueue_time = codel_get_time(); -+} -+ -+static inline u32 codel_time_to_us(codel_time_t val) -+{ -+ u64 valns = ((u64)val << CODEL_SHIFT); -+ -+ do_div(valns, NSEC_PER_USEC); -+ return (u32)valns; -+} -+ -+/** -+ * struct codel_params - contains codel parameters -+ * @target: target queue size (in time units) -+ * @interval: width of moving time window -+ * @ecn: is Explicit Congestion Notification enabled -+ */ -+struct codel_params { -+ codel_time_t target; -+ codel_time_t interval; -+ bool ecn; -+}; -+ -+/** -+ * struct codel_vars - contains codel variables -+ * @count: how many drops we've done since the last time we -+ * entered dropping state -+ * @lastcount: count at entry to dropping state -+ * @dropping: set to true if in dropping state -+ * @first_above_time: when we went (or will go) continuously above target -+ * for interval -+ * @drop_next: time to drop next packet, or when we dropped last -+ * @ldelay: sojourn time of last dequeued packet -+ */ -+struct codel_vars { -+ u32 count; -+ u32 lastcount; -+ bool dropping; -+ codel_time_t first_above_time; -+ codel_time_t drop_next; -+ codel_time_t ldelay; -+}; -+ -+/** -+ * struct codel_stats - contains codel shared variables and stats -+ * @maxpacket: largest packet we've seen so far -+ * @drop_count: temp count of dropped packets in dequeue() -+ * ecn_mark: number of packets we ECN marked instead of dropping -+ */ -+struct codel_stats { -+ u32 maxpacket; -+ u32 drop_count; -+ u32 ecn_mark; -+}; -+ -+static void codel_params_init(struct codel_params *params) -+{ -+ params->interval = MS2TIME(100); -+ params->target = MS2TIME(5); -+ params->ecn = false; -+} -+ -+static void codel_vars_init(struct codel_vars *vars) -+{ -+ vars->drop_next = 0; -+ vars->first_above_time = 0; -+ vars->dropping = false; /* exit dropping state */ -+ vars->count = 0; -+ vars->lastcount = 0; -+} -+ -+static void codel_stats_init(struct codel_stats *stats) -+{ -+ stats->maxpacket = 256; -+} -+ -+/* return interval/sqrt(x) with good precision -+ * relies on int_sqrt(unsigned long x) kernel implementation -+ */ -+static u32 codel_inv_sqrt(u32 _interval, u32 _x) -+{ -+ u64 interval = _interval; -+ unsigned long x = _x; -+ -+ /* Scale operands for max precision */ -+ -+#if BITS_PER_LONG == 64 -+ x <<= 32; /* On 64bit arches, we can prescale x by 32bits */ -+ interval <<= 16; -+#endif -+ -+ while (x < (1UL << (BITS_PER_LONG - 2))) { -+ x <<= 2; -+ interval <<= 1; -+ } -+ do_div(interval, int_sqrt(x)); -+ return (u32)interval; -+} -+ -+static codel_time_t codel_control_law(codel_time_t t, -+ codel_time_t interval, -+ u32 count) -+{ -+ return t + codel_inv_sqrt(interval, count); -+} -+ -+ -+static bool codel_should_drop(struct sk_buff *skb, -+ unsigned int *backlog, -+ struct codel_vars *vars, -+ struct codel_params *params, -+ struct codel_stats *stats, -+ codel_time_t now) -+{ -+ bool ok_to_drop; -+ -+ if (!skb) { -+ vars->first_above_time = 0; -+ return false; -+ } -+ -+ vars->ldelay = now - codel_get_enqueue_time(skb); -+ *backlog -= qdisc_pkt_len(skb); -+ -+ if (unlikely(qdisc_pkt_len(skb) > stats->maxpacket)) -+ stats->maxpacket = qdisc_pkt_len(skb); -+ -+ if (codel_time_before(vars->ldelay, params->target) || -+ *backlog <= stats->maxpacket) { -+ /* went below - stay below for at least interval */ -+ vars->first_above_time = 0; -+ return false; -+ } -+ ok_to_drop = false; -+ if (vars->first_above_time == 0) { -+ /* just went above from below. If we stay above -+ * for at least interval we'll say it's ok to drop -+ */ -+ vars->first_above_time = now + params->interval; -+ } else if (codel_time_after(now, vars->first_above_time)) { -+ ok_to_drop = true; -+ } -+ return ok_to_drop; -+} -+ -+typedef struct sk_buff * (*codel_skb_dequeue_t)(struct codel_vars *vars, -+ struct Qdisc *sch); -+ -+static struct sk_buff *codel_dequeue(struct Qdisc *sch, -+ struct codel_params *params, -+ struct codel_vars *vars, -+ struct codel_stats *stats, -+ codel_skb_dequeue_t dequeue_func, -+ u32 *backlog) -+{ -+ struct sk_buff *skb = dequeue_func(vars, sch); -+ codel_time_t now; -+ bool drop; -+ -+ if (!skb) { -+ vars->dropping = false; -+ return skb; -+ } -+ now = codel_get_time(); -+ drop = codel_should_drop(skb, backlog, vars, params, stats, now); -+ if (vars->dropping) { -+ if (!drop) { -+ /* sojourn time below target - leave dropping state */ -+ vars->dropping = false; -+ } else if (codel_time_after_eq(now, vars->drop_next)) { -+ /* It's time for the next drop. Drop the current -+ * packet and dequeue the next. The dequeue might -+ * take us out of dropping state. -+ * If not, schedule the next drop. -+ * A large backlog might result in drop rates so high -+ * that the next drop should happen now, -+ * hence the while loop. -+ */ -+ while (vars->dropping && -+ codel_time_after_eq(now, vars->drop_next)) { -+ if (++vars->count == 0) /* avoid zero divides */ -+ vars->count = ~0U; -+ if (params->ecn && INET_ECN_set_ce(skb)) { -+ stats->ecn_mark++; -+ vars->drop_next = -+ codel_control_law(vars->drop_next, -+ params->interval, -+ vars->count); -+ goto end; -+ } -+ qdisc_drop(skb, sch); -+ stats->drop_count++; -+ skb = dequeue_func(vars, sch); -+ if (!codel_should_drop(skb, backlog, -+ vars, params, stats, now)) { -+ /* leave dropping state */ -+ vars->dropping = false; -+ } else { -+ /* and schedule the next drop */ -+ vars->drop_next = -+ codel_control_law(vars->drop_next, -+ params->interval, -+ vars->count); -+ } -+ } -+ } -+ } else if (drop) { -+ if (params->ecn && INET_ECN_set_ce(skb)) { -+ stats->ecn_mark++; -+ } else { -+ qdisc_drop(skb, sch); -+ stats->drop_count++; -+ -+ skb = dequeue_func(vars, sch); -+ drop = codel_should_drop(skb, backlog, vars, params, -+ stats, now); -+ } -+ vars->dropping = true; -+ /* if min went above target close to when we last went below it -+ * assume that the drop rate that controlled the queue on the -+ * last cycle is a good starting point to control it now. -+ */ -+ if (codel_time_before(now - vars->drop_next, -+ 16 * params->interval)) { -+ vars->count = (vars->count - vars->lastcount) | 1; -+ } else { -+ vars->count = 1; -+ } -+ vars->lastcount = vars->count; -+ vars->drop_next = codel_control_law(now, params->interval, -+ vars->count); -+ } -+end: -+ return skb; -+} -+#endif -diff --git a/net/sched/Kconfig b/net/sched/Kconfig -index 75b58f8..fadd252 100644 ---- a/net/sched/Kconfig -+++ b/net/sched/Kconfig -@@ -250,6 +250,17 @@ config NET_SCH_QFQ - - If unsure, say N. - -+config NET_SCH_CODEL -+ tristate "Controlled Delay AQM (CODEL)" -+ help -+ Say Y here if you want to use the Controlled Delay (CODEL) -+ packet scheduling algorithm. -+ -+ To compile this driver as a module, choose M here: the module -+ will be called sch_codel. -+ -+ If unsure, say N. -+ - config NET_SCH_INGRESS - tristate "Ingress Qdisc" - depends on NET_CLS_ACT -diff --git a/net/sched/Makefile b/net/sched/Makefile -index 8cdf4e2..30fab03 100644 ---- a/net/sched/Makefile -+++ b/net/sched/Makefile -@@ -37,6 +37,7 @@ obj-$(CONFIG_NET_SCH_PLUG) += sch_plug.o - obj-$(CONFIG_NET_SCH_MQPRIO) += sch_mqprio.o - obj-$(CONFIG_NET_SCH_CHOKE) += sch_choke.o - obj-$(CONFIG_NET_SCH_QFQ) += sch_qfq.o -+obj-$(CONFIG_NET_SCH_CODEL) += sch_codel.o - - obj-$(CONFIG_NET_CLS_U32) += cls_u32.o - obj-$(CONFIG_NET_CLS_ROUTE4) += cls_route.o -diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c -new file mode 100644 -index 0000000..b4a1a81 ---- /dev/null -+++ b/net/sched/sch_codel.c -@@ -0,0 +1,275 @@ -+/* -+ * Codel - The Controlled-Delay Active Queue Management algorithm -+ * -+ * Copyright (C) 2011-2012 Kathleen Nichols <nichols(a)pollere.com> -+ * Copyright (C) 2011-2012 Van Jacobson <van(a)pollere.net> -+ * -+ * Implemented on linux by : -+ * Copyright (C) 2012 Michael D. Taht <dave.taht(a)bufferbloat.net> -+ * Copyright (C) 2012 Eric Dumazet <edumazet(a)google.com> -+ * -+ * Redistribution and use in source and binary forms, with or without -+ * modification, are permitted provided that the following conditions -+ * are met: -+ * 1. Redistributions of source code must retain the above copyright -+ * notice, this list of conditions, and the following disclaimer, -+ * without modification. -+ * 2. Redistributions in binary form must reproduce the above copyright -+ * notice, this list of conditions and the following disclaimer in the -+ * documentation and/or other materials provided with the distribution. -+ * 3. The names of the authors may not be used to endorse or promote products -+ * derived from this software without specific prior written permission. -+ * -+ * Alternatively, provided that this notice is retained in full, this -+ * software may be distributed under the terms of the GNU General -+ * Public License ("GPL") version 2, in which case the provisions of the -+ * GPL apply INSTEAD OF those given above. -+ * -+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS -+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR -+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT -+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, -+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT -+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, -+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY -+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT -+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH -+ * DAMAGE. -+ * -+ */ -+ -+#include <linux/module.h> -+#include <linux/slab.h> -+#include <linux/types.h> -+#include <linux/kernel.h> -+#include <linux/errno.h> -+#include <linux/skbuff.h> -+#include <net/pkt_sched.h> -+#include <net/codel.h> -+ -+ -+#define DEFAULT_CODEL_LIMIT 1000 -+ -+struct codel_sched_data { -+ struct codel_params params; -+ struct codel_vars vars; -+ struct codel_stats stats; -+ u32 drop_overlimit; -+}; -+ -+/* This is the specific function called from codel_dequeue() -+ * to dequeue a packet from queue. Note: backlog is handled in -+ * codel, we dont need to reduce it here. -+ */ -+static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch) -+{ -+ struct sk_buff *skb = __skb_dequeue(&sch->q); -+ -+ prefetch(&skb->end); /* we'll need skb_shinfo() */ -+ return skb; -+} -+ -+static struct sk_buff *codel_qdisc_dequeue(struct Qdisc *sch) -+{ -+ struct codel_sched_data *q = qdisc_priv(sch); -+ struct sk_buff *skb; -+ -+ skb = codel_dequeue(sch, &q->params, &q->vars, &q->stats, -+ dequeue, &sch->qstats.backlog); -+ /* We cant call qdisc_tree_decrease_qlen() if our qlen is 0, -+ * or HTB crashes. Defer it for next round. -+ */ -+ if (q->stats.drop_count && sch->q.qlen) { -+ qdisc_tree_decrease_qlen(sch, q->stats.drop_count); -+ q->stats.drop_count = 0; -+ } -+ if (skb) -+ qdisc_bstats_update(sch, skb); -+ return skb; -+} -+ -+static int codel_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch) -+{ -+ struct codel_sched_data *q; -+ -+ if (likely(qdisc_qlen(sch) < sch->limit)) { -+ codel_set_enqueue_time(skb); -+ return qdisc_enqueue_tail(skb, sch); -+ } -+ q = qdisc_priv(sch); -+ q->drop_overlimit++; -+ return qdisc_drop(skb, sch); -+} -+ -+static const struct nla_policy codel_policy[TCA_CODEL_MAX + 1] = { -+ [TCA_CODEL_TARGET] = { .type = NLA_U32 }, -+ [TCA_CODEL_LIMIT] = { .type = NLA_U32 }, -+ [TCA_CODEL_INTERVAL] = { .type = NLA_U32 }, -+ [TCA_CODEL_ECN] = { .type = NLA_U32 }, -+}; -+ -+static int codel_change(struct Qdisc *sch, struct nlattr *opt) -+{ -+ struct codel_sched_data *q = qdisc_priv(sch); -+ struct nlattr *tb[TCA_CODEL_MAX + 1]; -+ unsigned int qlen; -+ int err; -+ -+ if (!opt) -+ return -EINVAL; -+ -+ err = nla_parse_nested(tb, TCA_CODEL_MAX, opt, codel_policy); -+ if (err < 0) -+ return err; -+ -+ sch_tree_lock(sch); -+ -+ if (tb[TCA_CODEL_TARGET]) { -+ u32 target = nla_get_u32(tb[TCA_CODEL_TARGET]); -+ -+ q->params.target = ((u64)target * NSEC_PER_USEC) >> CODEL_SHIFT; -+ } -+ -+ if (tb[TCA_CODEL_INTERVAL]) { -+ u32 interval = nla_get_u32(tb[TCA_CODEL_INTERVAL]); -+ -+ q->params.interval = ((u64)interval * NSEC_PER_USEC) >> CODEL_SHIFT; -+ } -+ -+ if (tb[TCA_CODEL_LIMIT]) -+ sch->limit = nla_get_u32(tb[TCA_CODEL_LIMIT]); -+ -+ if (tb[TCA_CODEL_ECN]) -+ q->params.ecn = !!nla_get_u32(tb[TCA_CODEL_ECN]); -+ -+ qlen = sch->q.qlen; -+ while (sch->q.qlen > sch->limit) { -+ struct sk_buff *skb = __skb_dequeue(&sch->q); -+ -+ sch->qstats.backlog -= qdisc_pkt_len(skb); -+ qdisc_drop(skb, sch); -+ } -+ qdisc_tree_decrease_qlen(sch, qlen - sch->q.qlen); -+ -+ sch_tree_unlock(sch); -+ return 0; -+} -+ -+static int codel_init(struct Qdisc *sch, struct nlattr *opt) -+{ -+ struct codel_sched_data *q = qdisc_priv(sch); -+ -+ sch->limit = DEFAULT_CODEL_LIMIT; -+ -+ codel_params_init(&q->params); -+ codel_vars_init(&q->vars); -+ codel_stats_init(&q->stats); -+ -+ if (opt) { -+ int err = codel_change(sch, opt); -+ -+ if (err) -+ return err; -+ } -+ -+ if (sch->limit >= 1) -+ sch->flags |= TCQ_F_CAN_BYPASS; -+ else -+ sch->flags &= ~TCQ_F_CAN_BYPASS; -+ -+ return 0; -+} -+ -+static int codel_dump(struct Qdisc *sch, struct sk_buff *skb) -+{ -+ struct codel_sched_data *q = qdisc_priv(sch); -+ struct nlattr *opts; -+ -+ opts = nla_nest_start(skb, TCA_OPTIONS); -+ if (opts == NULL) -+ goto nla_put_failure; -+ -+ if (nla_put_u32(skb, TCA_CODEL_TARGET, -+ codel_time_to_us(q->params.target)) || -+ nla_put_u32(skb, TCA_CODEL_LIMIT, -+ sch->limit) || -+ nla_put_u32(skb, TCA_CODEL_INTERVAL, -+ codel_time_to_us(q->params.interval)) || -+ nla_put_u32(skb, TCA_CODEL_ECN, -+ q->params.ecn)) -+ goto nla_put_failure; -+ -+ return nla_nest_end(skb, opts); -+ -+nla_put_failure: -+ nla_nest_cancel(skb, opts); -+ return -1; -+} -+ -+static int codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d) -+{ -+ const struct codel_sched_data *q = qdisc_priv(sch); -+ struct tc_codel_xstats st = { -+ .maxpacket = q->stats.maxpacket, -+ .count = q->vars.count, -+ .lastcount = q->vars.lastcount, -+ .drop_overlimit = q->drop_overlimit, -+ .ldelay = codel_time_to_us(q->vars.ldelay), -+ .dropping = q->vars.dropping, -+ .ecn_mark = q->stats.ecn_mark, -+ }; -+ -+ if (q->vars.dropping) { -+ codel_tdiff_t delta = q->vars.drop_next - codel_get_time(); -+ -+ if (delta >= 0) -+ st.drop_next = codel_time_to_us(delta); -+ else -+ st.drop_next = -codel_time_to_us(-delta); -+ } -+ -+ return gnet_stats_copy_app(d, &st, sizeof(st)); -+} -+ -+static void codel_reset(struct Qdisc *sch) -+{ -+ struct codel_sched_data *q = qdisc_priv(sch); -+ -+ qdisc_reset_queue(sch); -+ codel_vars_init(&q->vars); -+} -+ -+static struct Qdisc_ops codel_qdisc_ops __read_mostly = { -+ .id = "codel", -+ .priv_size = sizeof(struct codel_sched_data), -+ -+ .enqueue = codel_qdisc_enqueue, -+ .dequeue = codel_qdisc_dequeue, -+ .peek = qdisc_peek_dequeued, -+ .init = codel_init, -+ .reset = codel_reset, -+ .change = codel_change, -+ .dump = codel_dump, -+ .dump_stats = codel_dump_stats, -+ .owner = THIS_MODULE, -+}; -+ -+static int __init codel_module_init(void) -+{ -+ return register_qdisc(&codel_qdisc_ops); -+} -+ -+static void __exit codel_module_exit(void) -+{ -+ unregister_qdisc(&codel_qdisc_ops); -+} -+ -+module_init(codel_module_init) -+module_exit(codel_module_exit) -+ -+MODULE_DESCRIPTION("Controlled Delay queue discipline"); -+MODULE_AUTHOR("Dave Taht"); -+MODULE_AUTHOR("Eric Dumazet"); -+MODULE_LICENSE("Dual BSD/GPL"); --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch b/src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch deleted file mode 100644 index 954d050..0000000 --- a/src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch +++ /dev/null @@ -1,189 +0,0 @@ -From: Eric Dumazet <edumazet(a)google.com> -Date: Sat, 12 May 2012 03:32:13 +0000 -Subject: [PATCH 2/7] codel: use Newton method instead of sqrt() and divides - -commit 536edd67109df5e0cdb2c4ee759e9bade7976367 upstream. - -As Van pointed out, interval/sqrt(count) can be implemented using -multiplies only. - -http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_methods_for_reciprocal_square_roots - -This patch implements the Newton method and reciprocal divide. - -Total cost is 15 cycles instead of 120 on my Corei5 machine (64bit -kernel). - -There is a small 'error' for count values < 5, but we don't really care. - -I reuse a hole in struct codel_vars : - - pack the dropping boolean into one bit - - use 31bit to store the reciprocal value of sqrt(count). - -Suggested-by: Van Jacobson <van(a)pollere.net> -Signed-off-by: Eric Dumazet <edumazet(a)google.com> -Cc: Dave Taht <dave.taht(a)bufferbloat.net> -Cc: Kathleen Nichols <nichols(a)pollere.com> -Cc: Tom Herbert <therbert(a)google.com> -Cc: Matt Mathis <mattmathis(a)google.com> -Cc: Yuchung Cheng <ycheng(a)google.com> -Cc: Nandita Dukkipati <nanditad(a)google.com> -Cc: Stephen Hemminger <shemminger(a)vyatta.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> ---- - include/net/codel.h | 68 ++++++++++++++++++++++++++++----------------------- - 1 file changed, 37 insertions(+), 31 deletions(-) - -diff --git a/include/net/codel.h b/include/net/codel.h -index bce2cef..bd8747c 100644 ---- a/include/net/codel.h -+++ b/include/net/codel.h -@@ -46,6 +46,7 @@ - #include <linux/skbuff.h> - #include <net/pkt_sched.h> - #include <net/inet_ecn.h> -+#include <linux/reciprocal_div.h> - - /* Controlling Queue Delay (CoDel) algorithm - * ========================================= -@@ -123,6 +124,7 @@ struct codel_params { - * entered dropping state - * @lastcount: count at entry to dropping state - * @dropping: set to true if in dropping state -+ * @rec_inv_sqrt: reciprocal value of sqrt(count) >> 1 - * @first_above_time: when we went (or will go) continuously above target - * for interval - * @drop_next: time to drop next packet, or when we dropped last -@@ -131,7 +133,8 @@ struct codel_params { - struct codel_vars { - u32 count; - u32 lastcount; -- bool dropping; -+ bool dropping:1; -+ u32 rec_inv_sqrt:31; - codel_time_t first_above_time; - codel_time_t drop_next; - codel_time_t ldelay; -@@ -158,11 +161,7 @@ static void codel_params_init(struct codel_params *params) - - static void codel_vars_init(struct codel_vars *vars) - { -- vars->drop_next = 0; -- vars->first_above_time = 0; -- vars->dropping = false; /* exit dropping state */ -- vars->count = 0; -- vars->lastcount = 0; -+ memset(vars, 0, sizeof(*vars)); - } - - static void codel_stats_init(struct codel_stats *stats) -@@ -170,38 +169,37 @@ static void codel_stats_init(struct codel_stats *stats) - stats->maxpacket = 256; - } - --/* return interval/sqrt(x) with good precision -- * relies on int_sqrt(unsigned long x) kernel implementation -+/* -+ * http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_me… -+ * new_invsqrt = (invsqrt / 2) * (3 - count * invsqrt^2) -+ * -+ * Here, invsqrt is a fixed point number (< 1.0), 31bit mantissa) - */ --static u32 codel_inv_sqrt(u32 _interval, u32 _x) -+static void codel_Newton_step(struct codel_vars *vars) - { -- u64 interval = _interval; -- unsigned long x = _x; -+ u32 invsqrt = vars->rec_inv_sqrt; -+ u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 31; -+ u64 val = (3LL << 31) - ((u64)vars->count * invsqrt2); - -- /* Scale operands for max precision */ -- --#if BITS_PER_LONG == 64 -- x <<= 32; /* On 64bit arches, we can prescale x by 32bits */ -- interval <<= 16; --#endif -+ val = (val * invsqrt) >> 32; - -- while (x < (1UL << (BITS_PER_LONG - 2))) { -- x <<= 2; -- interval <<= 1; -- } -- do_div(interval, int_sqrt(x)); -- return (u32)interval; -+ vars->rec_inv_sqrt = val; - } - -+/* -+ * CoDel control_law is t + interval/sqrt(count) -+ * We maintain in rec_inv_sqrt the reciprocal value of sqrt(count) to avoid -+ * both sqrt() and divide operation. -+ */ - static codel_time_t codel_control_law(codel_time_t t, - codel_time_t interval, -- u32 count) -+ u32 rec_inv_sqrt) - { -- return t + codel_inv_sqrt(interval, count); -+ return t + reciprocal_divide(interval, rec_inv_sqrt << 1); - } - - --static bool codel_should_drop(struct sk_buff *skb, -+static bool codel_should_drop(const struct sk_buff *skb, - unsigned int *backlog, - struct codel_vars *vars, - struct codel_params *params, -@@ -274,14 +272,16 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - */ - while (vars->dropping && - codel_time_after_eq(now, vars->drop_next)) { -- if (++vars->count == 0) /* avoid zero divides */ -- vars->count = ~0U; -+ vars->count++; /* dont care of possible wrap -+ * since there is no more divide -+ */ -+ codel_Newton_step(vars); - if (params->ecn && INET_ECN_set_ce(skb)) { - stats->ecn_mark++; - vars->drop_next = - codel_control_law(vars->drop_next, - params->interval, -- vars->count); -+ vars->rec_inv_sqrt); - goto end; - } - qdisc_drop(skb, sch); -@@ -296,7 +296,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - vars->drop_next = - codel_control_law(vars->drop_next, - params->interval, -- vars->count); -+ vars->rec_inv_sqrt); - } - } - } -@@ -319,12 +319,18 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - if (codel_time_before(now - vars->drop_next, - 16 * params->interval)) { - vars->count = (vars->count - vars->lastcount) | 1; -+ /* we dont care if rec_inv_sqrt approximation -+ * is not very precise : -+ * Next Newton steps will correct it quadratically. -+ */ -+ codel_Newton_step(vars); - } else { - vars->count = 1; -+ vars->rec_inv_sqrt = 0x7fffffff; - } - vars->lastcount = vars->count; - vars->drop_next = codel_control_law(now, params->interval, -- vars->count); -+ vars->rec_inv_sqrt); - } - end: - return skb; --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch b/src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch deleted file mode 100644 index e1d2260..0000000 --- a/src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch +++ /dev/null @@ -1,850 +0,0 @@ -From: Eric Dumazet <edumazet(a)google.com> -Date: Fri, 11 May 2012 09:30:50 +0000 -Subject: [PATCH 3/7] fq_codel: Fair Queue Codel AQM - -commit 4b549a2ef4bef9965d97cbd992ba67930cd3e0fe upstream. - -Fair Queue Codel packet scheduler - -Principles : - -- Packets are classified (internal classifier or external) on flows. -- This is a Stochastic model (as we use a hash, several flows might - be hashed on same slot) -- Each flow has a CoDel managed queue. -- Flows are linked onto two (Round Robin) lists, - so that new flows have priority on old ones. - -- For a given flow, packets are not reordered (CoDel uses a FIFO) -- head drops only. -- ECN capability is on by default. -- Very low memory footprint (64 bytes per flow) - -tc qdisc ... fq_codel [ limit PACKETS ] [ flows number ] - [ target TIME ] [ interval TIME ] [ noecn ] - [ quantum BYTES ] - -defaults : 1024 flows, 10240 packets limit, quantum : device MTU - target : 5ms (CoDel default) - interval : 100ms (CoDel default) - -Impressive results on load : - -class htb 1:1 root leaf 10: prio 0 quantum 1514 rate 200000Kbit ceil 200000Kbit burst 1475b/8 mpu 0b overhead 0b cburst 1475b/8 mpu 0b overhead 0b level 0 - Sent 43304920109 bytes 33063109 pkt (dropped 0, overlimits 0 requeues 0) - rate 201691Kbit 28595pps backlog 0b 312p requeues 0 - lended: 33063109 borrowed: 0 giants: 0 - tokens: -912 ctokens: -912 - -class fq_codel 10:1735 parent 10: - (dropped 1292, overlimits 0 requeues 0) - backlog 15140b 10p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 7.1ms -class fq_codel 10:4524 parent 10: - (dropped 1291, overlimits 0 requeues 0) - backlog 16654b 11p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 7.1ms -class fq_codel 10:4e74 parent 10: - (dropped 1290, overlimits 0 requeues 0) - backlog 6056b 4p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 6.4ms dropping drop_next 92.0ms -class fq_codel 10:628a parent 10: - (dropped 1289, overlimits 0 requeues 0) - backlog 7570b 5p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 5.4ms dropping drop_next 90.9ms -class fq_codel 10:a4b3 parent 10: - (dropped 302, overlimits 0 requeues 0) - backlog 16654b 11p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 7.1ms -class fq_codel 10:c3c2 parent 10: - (dropped 1284, overlimits 0 requeues 0) - backlog 13626b 9p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 5.9ms -class fq_codel 10:d331 parent 10: - (dropped 299, overlimits 0 requeues 0) - backlog 15140b 10p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 7.0ms -class fq_codel 10:d526 parent 10: - (dropped 12160, overlimits 0 requeues 0) - backlog 35870b 211p requeues 0 - deficit 1508 count 12160 lastcount 1 ldelay 15.3ms dropping drop_next 247us -class fq_codel 10:e2c6 parent 10: - (dropped 1288, overlimits 0 requeues 0) - backlog 15140b 10p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 7.1ms -class fq_codel 10:eab5 parent 10: - (dropped 1285, overlimits 0 requeues 0) - backlog 16654b 11p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 5.9ms -class fq_codel 10:f220 parent 10: - (dropped 1289, overlimits 0 requeues 0) - backlog 15140b 10p requeues 0 - deficit 1514 count 1 lastcount 1 ldelay 7.1ms - -qdisc htb 1: root refcnt 6 r2q 10 default 1 direct_packets_stat 0 ver 3.17 - Sent 43331086547 bytes 33092812 pkt (dropped 0, overlimits 66063544 requeues 71) - rate 201697Kbit 28602pps backlog 0b 260p requeues 71 -qdisc fq_codel 10: parent 1:1 limit 10240p flows 65536 target 5.0ms interval 100.0ms ecn - Sent 43331086547 bytes 33092812 pkt (dropped 949359, overlimits 0 requeues 0) - rate 201697Kbit 28602pps backlog 189352b 260p requeues 0 - maxpacket 1514 drop_overlimit 0 new_flow_count 5582 ecn_mark 125593 - new_flows_len 0 old_flows_len 11 - -PING 172.30.42.18 (172.30.42.18) 56(84) bytes of data. -64 bytes from 172.30.42.18: icmp_req=1 ttl=64 time=0.227 ms -64 bytes from 172.30.42.18: icmp_req=2 ttl=64 time=0.165 ms -64 bytes from 172.30.42.18: icmp_req=3 ttl=64 time=0.166 ms -64 bytes from 172.30.42.18: icmp_req=4 ttl=64 time=0.151 ms -64 bytes from 172.30.42.18: icmp_req=5 ttl=64 time=0.164 ms -64 bytes from 172.30.42.18: icmp_req=6 ttl=64 time=0.172 ms -64 bytes from 172.30.42.18: icmp_req=7 ttl=64 time=0.175 ms -64 bytes from 172.30.42.18: icmp_req=8 ttl=64 time=0.183 ms -64 bytes from 172.30.42.18: icmp_req=9 ttl=64 time=0.158 ms -64 bytes from 172.30.42.18: icmp_req=10 ttl=64 time=0.200 ms - -10 packets transmitted, 10 received, 0% packet loss, time 8999ms -rtt min/avg/max/mdev = 0.151/0.176/0.227/0.022 ms - -Much better than SFQ because of priority given to new flows, and fast -path dirtying less cache lines. - -Signed-off-by: Eric Dumazet <edumazet(a)google.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> ---- - include/linux/pkt_sched.h | 54 ++++ - net/sched/Kconfig | 11 + - net/sched/Makefile | 1 + - net/sched/sch_fq_codel.c | 624 +++++++++++++++++++++++++++++++++++++++++++++ - 4 files changed, 690 insertions(+) - create mode 100644 net/sched/sch_fq_codel.c - -diff --git a/include/linux/pkt_sched.h b/include/linux/pkt_sched.h -index cde56c2..32aef0a 100644 ---- a/include/linux/pkt_sched.h -+++ b/include/linux/pkt_sched.h -@@ -681,4 +681,58 @@ struct tc_codel_xstats { - __u32 dropping; /* are we in dropping state ? */ - }; - -+/* FQ_CODEL */ -+ -+enum { -+ TCA_FQ_CODEL_UNSPEC, -+ TCA_FQ_CODEL_TARGET, -+ TCA_FQ_CODEL_LIMIT, -+ TCA_FQ_CODEL_INTERVAL, -+ TCA_FQ_CODEL_ECN, -+ TCA_FQ_CODEL_FLOWS, -+ TCA_FQ_CODEL_QUANTUM, -+ __TCA_FQ_CODEL_MAX -+}; -+ -+#define TCA_FQ_CODEL_MAX (__TCA_FQ_CODEL_MAX - 1) -+ -+enum { -+ TCA_FQ_CODEL_XSTATS_QDISC, -+ TCA_FQ_CODEL_XSTATS_CLASS, -+}; -+ -+struct tc_fq_codel_qd_stats { -+ __u32 maxpacket; /* largest packet we've seen so far */ -+ __u32 drop_overlimit; /* number of time max qdisc -+ * packet limit was hit -+ */ -+ __u32 ecn_mark; /* number of packets we ECN marked -+ * instead of being dropped -+ */ -+ __u32 new_flow_count; /* number of time packets -+ * created a 'new flow' -+ */ -+ __u32 new_flows_len; /* count of flows in new list */ -+ __u32 old_flows_len; /* count of flows in old list */ -+}; -+ -+struct tc_fq_codel_cl_stats { -+ __s32 deficit; -+ __u32 ldelay; /* in-queue delay seen by most recently -+ * dequeued packet -+ */ -+ __u32 count; -+ __u32 lastcount; -+ __u32 dropping; -+ __s32 drop_next; -+}; -+ -+struct tc_fq_codel_xstats { -+ __u32 type; -+ union { -+ struct tc_fq_codel_qd_stats qdisc_stats; -+ struct tc_fq_codel_cl_stats class_stats; -+ }; -+}; -+ - #endif -diff --git a/net/sched/Kconfig b/net/sched/Kconfig -index fadd252..e7a8976 100644 ---- a/net/sched/Kconfig -+++ b/net/sched/Kconfig -@@ -261,6 +261,17 @@ config NET_SCH_CODEL - - If unsure, say N. - -+config NET_SCH_FQ_CODEL -+ tristate "Fair Queue Controlled Delay AQM (FQ_CODEL)" -+ help -+ Say Y here if you want to use the FQ Controlled Delay (FQ_CODEL) -+ packet scheduling algorithm. -+ -+ To compile this driver as a module, choose M here: the module -+ will be called sch_fq_codel. -+ -+ If unsure, say N. -+ - config NET_SCH_INGRESS - tristate "Ingress Qdisc" - depends on NET_CLS_ACT -diff --git a/net/sched/Makefile b/net/sched/Makefile -index 30fab03..5940a19 100644 ---- a/net/sched/Makefile -+++ b/net/sched/Makefile -@@ -38,6 +38,7 @@ obj-$(CONFIG_NET_SCH_MQPRIO) += sch_mqprio.o - obj-$(CONFIG_NET_SCH_CHOKE) += sch_choke.o - obj-$(CONFIG_NET_SCH_QFQ) += sch_qfq.o - obj-$(CONFIG_NET_SCH_CODEL) += sch_codel.o -+obj-$(CONFIG_NET_SCH_FQ_CODEL) += sch_fq_codel.o - - obj-$(CONFIG_NET_CLS_U32) += cls_u32.o - obj-$(CONFIG_NET_CLS_ROUTE4) += cls_route.o -diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c -new file mode 100644 -index 0000000..a7b3754 ---- /dev/null -+++ b/net/sched/sch_fq_codel.c -@@ -0,0 +1,624 @@ -+/* -+ * Fair Queue CoDel 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. -+ * -+ * Copyright (C) 2012 Eric Dumazet <edumazet(a)google.com> -+ */ -+ -+#include <linux/module.h> -+#include <linux/types.h> -+#include <linux/kernel.h> -+#include <linux/jiffies.h> -+#include <linux/string.h> -+#include <linux/in.h> -+#include <linux/errno.h> -+#include <linux/init.h> -+#include <linux/skbuff.h> -+#include <linux/jhash.h> -+#include <linux/slab.h> -+#include <linux/vmalloc.h> -+#include <net/netlink.h> -+#include <net/pkt_sched.h> -+#include <net/flow_keys.h> -+#include <net/codel.h> -+ -+/* Fair Queue CoDel. -+ * -+ * Principles : -+ * Packets are classified (internal classifier or external) on flows. -+ * This is a Stochastic model (as we use a hash, several flows -+ * might be hashed on same slot) -+ * Each flow has a CoDel managed queue. -+ * Flows are linked onto two (Round Robin) lists, -+ * so that new flows have priority on old ones. -+ * -+ * For a given flow, packets are not reordered (CoDel uses a FIFO) -+ * head drops only. -+ * ECN capability is on by default. -+ * Low memory footprint (64 bytes per flow) -+ */ -+ -+struct fq_codel_flow { -+ struct sk_buff *head; -+ struct sk_buff *tail; -+ struct list_head flowchain; -+ int deficit; -+ u32 dropped; /* number of drops (or ECN marks) on this flow */ -+ struct codel_vars cvars; -+}; /* please try to keep this structure <= 64 bytes */ -+ -+struct fq_codel_sched_data { -+ struct tcf_proto *filter_list; /* optional external classifier */ -+ struct fq_codel_flow *flows; /* Flows table [flows_cnt] */ -+ u32 *backlogs; /* backlog table [flows_cnt] */ -+ u32 flows_cnt; /* number of flows */ -+ u32 perturbation; /* hash perturbation */ -+ u32 quantum; /* psched_mtu(qdisc_dev(sch)); */ -+ struct codel_params cparams; -+ struct codel_stats cstats; -+ u32 drop_overlimit; -+ u32 new_flow_count; -+ -+ struct list_head new_flows; /* list of new flows */ -+ struct list_head old_flows; /* list of old flows */ -+}; -+ -+static unsigned int fq_codel_hash(const struct fq_codel_sched_data *q, -+ const struct sk_buff *skb) -+{ -+ struct flow_keys keys; -+ unsigned int hash; -+ -+ skb_flow_dissect(skb, &keys); -+ hash = jhash_3words((__force u32)keys.dst, -+ (__force u32)keys.src ^ keys.ip_proto, -+ (__force u32)keys.ports, q->perturbation); -+ return ((u64)hash * q->flows_cnt) >> 32; -+} -+ -+static unsigned int fq_codel_classify(struct sk_buff *skb, struct Qdisc *sch, -+ int *qerr) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ struct tcf_result res; -+ int result; -+ -+ if (TC_H_MAJ(skb->priority) == sch->handle && -+ TC_H_MIN(skb->priority) > 0 && -+ TC_H_MIN(skb->priority) <= q->flows_cnt) -+ return TC_H_MIN(skb->priority); -+ -+ if (!q->filter_list) -+ return fq_codel_hash(q, skb) + 1; -+ -+ *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; -+ result = tc_classify(skb, q->filter_list, &res); -+ if (result >= 0) { -+#ifdef CONFIG_NET_CLS_ACT -+ switch (result) { -+ case TC_ACT_STOLEN: -+ case TC_ACT_QUEUED: -+ *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; -+ case TC_ACT_SHOT: -+ return 0; -+ } -+#endif -+ if (TC_H_MIN(res.classid) <= q->flows_cnt) -+ return TC_H_MIN(res.classid); -+ } -+ return 0; -+} -+ -+/* helper functions : might be changed when/if skb use a standard list_head */ -+ -+/* remove one skb from head of slot queue */ -+static inline struct sk_buff *dequeue_head(struct fq_codel_flow *flow) -+{ -+ struct sk_buff *skb = flow->head; -+ -+ flow->head = skb->next; -+ skb->next = NULL; -+ return skb; -+} -+ -+/* add skb to flow queue (tail add) */ -+static inline void flow_queue_add(struct fq_codel_flow *flow, -+ struct sk_buff *skb) -+{ -+ if (flow->head == NULL) -+ flow->head = skb; -+ else -+ flow->tail->next = skb; -+ flow->tail = skb; -+ skb->next = NULL; -+} -+ -+static unsigned int fq_codel_drop(struct Qdisc *sch) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ struct sk_buff *skb; -+ unsigned int maxbacklog = 0, idx = 0, i, len; -+ struct fq_codel_flow *flow; -+ -+ /* Queue is full! Find the fat flow and drop packet from it. -+ * This might sound expensive, but with 1024 flows, we scan -+ * 4KB of memory, and we dont need to handle a complex tree -+ * in fast path (packet queue/enqueue) with many cache misses. -+ */ -+ for (i = 0; i < q->flows_cnt; i++) { -+ if (q->backlogs[i] > maxbacklog) { -+ maxbacklog = q->backlogs[i]; -+ idx = i; -+ } -+ } -+ flow = &q->flows[idx]; -+ skb = dequeue_head(flow); -+ len = qdisc_pkt_len(skb); -+ q->backlogs[idx] -= len; -+ kfree_skb(skb); -+ sch->q.qlen--; -+ sch->qstats.drops++; -+ sch->qstats.backlog -= len; -+ flow->dropped++; -+ return idx; -+} -+ -+static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ unsigned int idx; -+ struct fq_codel_flow *flow; -+ int uninitialized_var(ret); -+ -+ idx = fq_codel_classify(skb, sch, &ret); -+ if (idx == 0) { -+ if (ret & __NET_XMIT_BYPASS) -+ sch->qstats.drops++; -+ kfree_skb(skb); -+ return ret; -+ } -+ idx--; -+ -+ codel_set_enqueue_time(skb); -+ flow = &q->flows[idx]; -+ flow_queue_add(flow, skb); -+ q->backlogs[idx] += qdisc_pkt_len(skb); -+ sch->qstats.backlog += qdisc_pkt_len(skb); -+ -+ if (list_empty(&flow->flowchain)) { -+ list_add_tail(&flow->flowchain, &q->new_flows); -+ codel_vars_init(&flow->cvars); -+ q->new_flow_count++; -+ flow->deficit = q->quantum; -+ flow->dropped = 0; -+ } -+ if (++sch->q.qlen < sch->limit) -+ return NET_XMIT_SUCCESS; -+ -+ q->drop_overlimit++; -+ /* Return Congestion Notification only if we dropped a packet -+ * from this flow. -+ */ -+ if (fq_codel_drop(sch) == idx) -+ return NET_XMIT_CN; -+ -+ /* As we dropped a packet, better let upper stack know this */ -+ qdisc_tree_decrease_qlen(sch, 1); -+ return NET_XMIT_SUCCESS; -+} -+ -+/* This is the specific function called from codel_dequeue() -+ * to dequeue a packet from queue. Note: backlog is handled in -+ * codel, we dont need to reduce it here. -+ */ -+static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch) -+{ -+ struct fq_codel_flow *flow; -+ struct sk_buff *skb = NULL; -+ -+ flow = container_of(vars, struct fq_codel_flow, cvars); -+ if (flow->head) { -+ skb = dequeue_head(flow); -+ sch->qstats.backlog -= qdisc_pkt_len(skb); -+ sch->q.qlen--; -+ } -+ return skb; -+} -+ -+static struct sk_buff *fq_codel_dequeue(struct Qdisc *sch) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ struct sk_buff *skb; -+ struct fq_codel_flow *flow; -+ struct list_head *head; -+ u32 prev_drop_count, prev_ecn_mark; -+ -+begin: -+ head = &q->new_flows; -+ if (list_empty(head)) { -+ head = &q->old_flows; -+ if (list_empty(head)) -+ return NULL; -+ } -+ flow = list_first_entry(head, struct fq_codel_flow, flowchain); -+ -+ if (flow->deficit <= 0) { -+ flow->deficit += q->quantum; -+ list_move_tail(&flow->flowchain, &q->old_flows); -+ goto begin; -+ } -+ -+ prev_drop_count = q->cstats.drop_count; -+ prev_ecn_mark = q->cstats.ecn_mark; -+ -+ skb = codel_dequeue(sch, &q->cparams, &flow->cvars, &q->cstats, -+ dequeue, &q->backlogs[flow - q->flows]); -+ -+ flow->dropped += q->cstats.drop_count - prev_drop_count; -+ flow->dropped += q->cstats.ecn_mark - prev_ecn_mark; -+ -+ if (!skb) { -+ /* force a pass through old_flows to prevent starvation */ -+ if ((head == &q->new_flows) && !list_empty(&q->old_flows)) -+ list_move_tail(&flow->flowchain, &q->old_flows); -+ else -+ list_del_init(&flow->flowchain); -+ goto begin; -+ } -+ qdisc_bstats_update(sch, skb); -+ flow->deficit -= qdisc_pkt_len(skb); -+ /* We cant call qdisc_tree_decrease_qlen() if our qlen is 0, -+ * or HTB crashes. Defer it for next round. -+ */ -+ if (q->cstats.drop_count && sch->q.qlen) { -+ qdisc_tree_decrease_qlen(sch, q->cstats.drop_count); -+ q->cstats.drop_count = 0; -+ } -+ return skb; -+} -+ -+static void fq_codel_reset(struct Qdisc *sch) -+{ -+ struct sk_buff *skb; -+ -+ while ((skb = fq_codel_dequeue(sch)) != NULL) -+ kfree_skb(skb); -+} -+ -+static const struct nla_policy fq_codel_policy[TCA_FQ_CODEL_MAX + 1] = { -+ [TCA_FQ_CODEL_TARGET] = { .type = NLA_U32 }, -+ [TCA_FQ_CODEL_LIMIT] = { .type = NLA_U32 }, -+ [TCA_FQ_CODEL_INTERVAL] = { .type = NLA_U32 }, -+ [TCA_FQ_CODEL_ECN] = { .type = NLA_U32 }, -+ [TCA_FQ_CODEL_FLOWS] = { .type = NLA_U32 }, -+ [TCA_FQ_CODEL_QUANTUM] = { .type = NLA_U32 }, -+}; -+ -+static int fq_codel_change(struct Qdisc *sch, struct nlattr *opt) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ struct nlattr *tb[TCA_FQ_CODEL_MAX + 1]; -+ int err; -+ -+ if (!opt) -+ return -EINVAL; -+ -+ err = nla_parse_nested(tb, TCA_FQ_CODEL_MAX, opt, fq_codel_policy); -+ if (err < 0) -+ return err; -+ if (tb[TCA_FQ_CODEL_FLOWS]) { -+ if (q->flows) -+ return -EINVAL; -+ q->flows_cnt = nla_get_u32(tb[TCA_FQ_CODEL_FLOWS]); -+ if (!q->flows_cnt || -+ q->flows_cnt > 65536) -+ return -EINVAL; -+ } -+ sch_tree_lock(sch); -+ -+ if (tb[TCA_FQ_CODEL_TARGET]) { -+ u64 target = nla_get_u32(tb[TCA_FQ_CODEL_TARGET]); -+ -+ q->cparams.target = (target * NSEC_PER_USEC) >> CODEL_SHIFT; -+ } -+ -+ if (tb[TCA_FQ_CODEL_INTERVAL]) { -+ u64 interval = nla_get_u32(tb[TCA_FQ_CODEL_INTERVAL]); -+ -+ q->cparams.interval = (interval * NSEC_PER_USEC) >> CODEL_SHIFT; -+ } -+ -+ if (tb[TCA_FQ_CODEL_LIMIT]) -+ sch->limit = nla_get_u32(tb[TCA_FQ_CODEL_LIMIT]); -+ -+ if (tb[TCA_FQ_CODEL_ECN]) -+ q->cparams.ecn = !!nla_get_u32(tb[TCA_FQ_CODEL_ECN]); -+ -+ if (tb[TCA_FQ_CODEL_QUANTUM]) -+ q->quantum = max(256U, nla_get_u32(tb[TCA_FQ_CODEL_QUANTUM])); -+ -+ while (sch->q.qlen > sch->limit) { -+ struct sk_buff *skb = fq_codel_dequeue(sch); -+ -+ kfree_skb(skb); -+ q->cstats.drop_count++; -+ } -+ qdisc_tree_decrease_qlen(sch, q->cstats.drop_count); -+ q->cstats.drop_count = 0; -+ -+ sch_tree_unlock(sch); -+ return 0; -+} -+ -+static void *fq_codel_zalloc(size_t sz) -+{ -+ void *ptr = kzalloc(sz, GFP_KERNEL | __GFP_NOWARN); -+ -+ if (!ptr) -+ ptr = vzalloc(sz); -+ return ptr; -+} -+ -+static void fq_codel_free(void *addr) -+{ -+ if (addr) { -+ if (is_vmalloc_addr(addr)) -+ vfree(addr); -+ else -+ kfree(addr); -+ } -+} -+ -+static void fq_codel_destroy(struct Qdisc *sch) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ -+ tcf_destroy_chain(&q->filter_list); -+ fq_codel_free(q->backlogs); -+ fq_codel_free(q->flows); -+} -+ -+static int fq_codel_init(struct Qdisc *sch, struct nlattr *opt) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ int i; -+ -+ sch->limit = 10*1024; -+ q->flows_cnt = 1024; -+ q->quantum = psched_mtu(qdisc_dev(sch)); -+ q->perturbation = net_random(); -+ INIT_LIST_HEAD(&q->new_flows); -+ INIT_LIST_HEAD(&q->old_flows); -+ codel_params_init(&q->cparams); -+ codel_stats_init(&q->cstats); -+ q->cparams.ecn = true; -+ -+ if (opt) { -+ int err = fq_codel_change(sch, opt); -+ if (err) -+ return err; -+ } -+ -+ if (!q->flows) { -+ q->flows = fq_codel_zalloc(q->flows_cnt * -+ sizeof(struct fq_codel_flow)); -+ if (!q->flows) -+ return -ENOMEM; -+ q->backlogs = fq_codel_zalloc(q->flows_cnt * sizeof(u32)); -+ if (!q->backlogs) { -+ fq_codel_free(q->flows); -+ return -ENOMEM; -+ } -+ for (i = 0; i < q->flows_cnt; i++) { -+ struct fq_codel_flow *flow = q->flows + i; -+ -+ INIT_LIST_HEAD(&flow->flowchain); -+ } -+ } -+ if (sch->limit >= 1) -+ sch->flags |= TCQ_F_CAN_BYPASS; -+ else -+ sch->flags &= ~TCQ_F_CAN_BYPASS; -+ return 0; -+} -+ -+static int fq_codel_dump(struct Qdisc *sch, struct sk_buff *skb) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ struct nlattr *opts; -+ -+ opts = nla_nest_start(skb, TCA_OPTIONS); -+ if (opts == NULL) -+ goto nla_put_failure; -+ -+ if (nla_put_u32(skb, TCA_FQ_CODEL_TARGET, -+ codel_time_to_us(q->cparams.target)) || -+ nla_put_u32(skb, TCA_FQ_CODEL_LIMIT, -+ sch->limit) || -+ nla_put_u32(skb, TCA_FQ_CODEL_INTERVAL, -+ codel_time_to_us(q->cparams.interval)) || -+ nla_put_u32(skb, TCA_FQ_CODEL_ECN, -+ q->cparams.ecn) || -+ nla_put_u32(skb, TCA_FQ_CODEL_QUANTUM, -+ q->quantum) || -+ nla_put_u32(skb, TCA_FQ_CODEL_FLOWS, -+ q->flows_cnt)) -+ goto nla_put_failure; -+ -+ nla_nest_end(skb, opts); -+ return skb->len; -+ -+nla_put_failure: -+ return -1; -+} -+ -+static int fq_codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ struct tc_fq_codel_xstats st = { -+ .type = TCA_FQ_CODEL_XSTATS_QDISC, -+ .qdisc_stats.maxpacket = q->cstats.maxpacket, -+ .qdisc_stats.drop_overlimit = q->drop_overlimit, -+ .qdisc_stats.ecn_mark = q->cstats.ecn_mark, -+ .qdisc_stats.new_flow_count = q->new_flow_count, -+ }; -+ struct list_head *pos; -+ -+ list_for_each(pos, &q->new_flows) -+ st.qdisc_stats.new_flows_len++; -+ -+ list_for_each(pos, &q->old_flows) -+ st.qdisc_stats.old_flows_len++; -+ -+ return gnet_stats_copy_app(d, &st, sizeof(st)); -+} -+ -+static struct Qdisc *fq_codel_leaf(struct Qdisc *sch, unsigned long arg) -+{ -+ return NULL; -+} -+ -+static unsigned long fq_codel_get(struct Qdisc *sch, u32 classid) -+{ -+ return 0; -+} -+ -+static unsigned long fq_codel_bind(struct Qdisc *sch, unsigned long parent, -+ u32 classid) -+{ -+ /* we cannot bypass queue discipline anymore */ -+ sch->flags &= ~TCQ_F_CAN_BYPASS; -+ return 0; -+} -+ -+static void fq_codel_put(struct Qdisc *q, unsigned long cl) -+{ -+} -+ -+static struct tcf_proto **fq_codel_find_tcf(struct Qdisc *sch, unsigned long cl) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ -+ if (cl) -+ return NULL; -+ return &q->filter_list; -+} -+ -+static int fq_codel_dump_class(struct Qdisc *sch, unsigned long cl, -+ struct sk_buff *skb, struct tcmsg *tcm) -+{ -+ tcm->tcm_handle |= TC_H_MIN(cl); -+ return 0; -+} -+ -+static int fq_codel_dump_class_stats(struct Qdisc *sch, unsigned long cl, -+ struct gnet_dump *d) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ u32 idx = cl - 1; -+ struct gnet_stats_queue qs = { 0 }; -+ struct tc_fq_codel_xstats xstats; -+ -+ if (idx < q->flows_cnt) { -+ const struct fq_codel_flow *flow = &q->flows[idx]; -+ const struct sk_buff *skb = flow->head; -+ -+ memset(&xstats, 0, sizeof(xstats)); -+ xstats.type = TCA_FQ_CODEL_XSTATS_CLASS; -+ xstats.class_stats.deficit = flow->deficit; -+ xstats.class_stats.ldelay = -+ codel_time_to_us(flow->cvars.ldelay); -+ xstats.class_stats.count = flow->cvars.count; -+ xstats.class_stats.lastcount = flow->cvars.lastcount; -+ xstats.class_stats.dropping = flow->cvars.dropping; -+ if (flow->cvars.dropping) { -+ codel_tdiff_t delta = flow->cvars.drop_next - -+ codel_get_time(); -+ -+ xstats.class_stats.drop_next = (delta >= 0) ? -+ codel_time_to_us(delta) : -+ -codel_time_to_us(-delta); -+ } -+ while (skb) { -+ qs.qlen++; -+ skb = skb->next; -+ } -+ qs.backlog = q->backlogs[idx]; -+ qs.drops = flow->dropped; -+ } -+ if (gnet_stats_copy_queue(d, &qs) < 0) -+ return -1; -+ if (idx < q->flows_cnt) -+ return gnet_stats_copy_app(d, &xstats, sizeof(xstats)); -+ return 0; -+} -+ -+static void fq_codel_walk(struct Qdisc *sch, struct qdisc_walker *arg) -+{ -+ struct fq_codel_sched_data *q = qdisc_priv(sch); -+ unsigned int i; -+ -+ if (arg->stop) -+ return; -+ -+ for (i = 0; i < q->flows_cnt; i++) { -+ if (list_empty(&q->flows[i].flowchain) || -+ arg->count < arg->skip) { -+ arg->count++; -+ continue; -+ } -+ if (arg->fn(sch, i + 1, arg) < 0) { -+ arg->stop = 1; -+ break; -+ } -+ arg->count++; -+ } -+} -+ -+static const struct Qdisc_class_ops fq_codel_class_ops = { -+ .leaf = fq_codel_leaf, -+ .get = fq_codel_get, -+ .put = fq_codel_put, -+ .tcf_chain = fq_codel_find_tcf, -+ .bind_tcf = fq_codel_bind, -+ .unbind_tcf = fq_codel_put, -+ .dump = fq_codel_dump_class, -+ .dump_stats = fq_codel_dump_class_stats, -+ .walk = fq_codel_walk, -+}; -+ -+static struct Qdisc_ops fq_codel_qdisc_ops __read_mostly = { -+ .cl_ops = &fq_codel_class_ops, -+ .id = "fq_codel", -+ .priv_size = sizeof(struct fq_codel_sched_data), -+ .enqueue = fq_codel_enqueue, -+ .dequeue = fq_codel_dequeue, -+ .peek = qdisc_peek_dequeued, -+ .drop = fq_codel_drop, -+ .init = fq_codel_init, -+ .reset = fq_codel_reset, -+ .destroy = fq_codel_destroy, -+ .change = fq_codel_change, -+ .dump = fq_codel_dump, -+ .dump_stats = fq_codel_dump_stats, -+ .owner = THIS_MODULE, -+}; -+ -+static int __init fq_codel_module_init(void) -+{ -+ return register_qdisc(&fq_codel_qdisc_ops); -+} -+ -+static void __exit fq_codel_module_exit(void) -+{ -+ unregister_qdisc(&fq_codel_qdisc_ops); -+} -+ -+module_init(fq_codel_module_init) -+module_exit(fq_codel_module_exit) -+MODULE_AUTHOR("Eric Dumazet"); -+MODULE_LICENSE("GPL"); --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch b/src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch deleted file mode 100644 index 0d3a886..0000000 --- a/src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch +++ /dev/null @@ -1,37 +0,0 @@ -From: Geert Uytterhoeven <geert(a)linux-m68k.org> -Date: Mon, 14 May 2012 09:47:05 +0000 -Subject: [PATCH 4/7] net/codel: Add missing #include <linux/prefetch.h> -MIME-Version: 1.0 -Content-Type: text/plain; charset=UTF-8 -Content-Transfer-Encoding: 8bit - -commit ce5b4b977127ee20c3f9c3fd3637cd3796f649f5 upstream. - -m68k allmodconfig: - -net/sched/sch_codel.c: In function ‘dequeue’: -net/sched/sch_codel.c:70: error: implicit declaration of function ‘prefetch’ -make[1]: *** [net/sched/sch_codel.o] Error 1 - -Signed-off-by: Geert Uytterhoeven <geert(a)linux-m68k.org> -Acked-by: Eric Dumazet <edumazet(a)google.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> ---- - net/sched/sch_codel.c | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c -index b4a1a81..213ef60 100644 ---- a/net/sched/sch_codel.c -+++ b/net/sched/sch_codel.c -@@ -46,6 +46,7 @@ - #include <linux/kernel.h> - #include <linux/errno.h> - #include <linux/skbuff.h> -+#include <linux/prefetch.h> - #include <net/pkt_sched.h> - #include <net/codel.h> - --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch b/src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch deleted file mode 100644 index 728c12c..0000000 --- a/src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch +++ /dev/null @@ -1,55 +0,0 @@ -From: Sasha Levin <levinsasha928(a)gmail.com> -Date: Mon, 14 May 2012 11:57:06 +0000 -Subject: [PATCH 5/7] net: codel: fix build errors - -commit 669d67bf777def468970f2dcba1537edf3b2d329 upstream. - -Fix the following build error: - -net/sched/sch_fq_codel.c: In function 'fq_codel_dump_stats': -net/sched/sch_fq_codel.c:464:3: error: unknown field 'qdisc_stats' specified in initializer -net/sched/sch_fq_codel.c:464:3: warning: missing braces around initializer -net/sched/sch_fq_codel.c:464:3: warning: (near initialization for 'st.<anonymous>') -net/sched/sch_fq_codel.c:465:3: error: unknown field 'qdisc_stats' specified in initializer -net/sched/sch_fq_codel.c:465:3: warning: excess elements in struct initializer -net/sched/sch_fq_codel.c:465:3: warning: (near initialization for 'st') -net/sched/sch_fq_codel.c:466:3: error: unknown field 'qdisc_stats' specified in initializer -net/sched/sch_fq_codel.c:466:3: warning: excess elements in struct initializer -net/sched/sch_fq_codel.c:466:3: warning: (near initialization for 'st') -net/sched/sch_fq_codel.c:467:3: error: unknown field 'qdisc_stats' specified in initializer -net/sched/sch_fq_codel.c:467:3: warning: excess elements in struct initializer -net/sched/sch_fq_codel.c:467:3: warning: (near initialization for 'st') -make[1]: *** [net/sched/sch_fq_codel.o] Error 1 - -Signed-off-by: Sasha Levin <levinsasha928(a)gmail.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> ---- - net/sched/sch_fq_codel.c | 9 +++++---- - 1 file changed, 5 insertions(+), 4 deletions(-) - -diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c -index a7b3754..337ff20 100644 ---- a/net/sched/sch_fq_codel.c -+++ b/net/sched/sch_fq_codel.c -@@ -461,13 +461,14 @@ static int fq_codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d) - struct fq_codel_sched_data *q = qdisc_priv(sch); - struct tc_fq_codel_xstats st = { - .type = TCA_FQ_CODEL_XSTATS_QDISC, -- .qdisc_stats.maxpacket = q->cstats.maxpacket, -- .qdisc_stats.drop_overlimit = q->drop_overlimit, -- .qdisc_stats.ecn_mark = q->cstats.ecn_mark, -- .qdisc_stats.new_flow_count = q->new_flow_count, - }; - struct list_head *pos; - -+ st.qdisc_stats.maxpacket = q->cstats.maxpacket; -+ st.qdisc_stats.drop_overlimit = q->drop_overlimit; -+ st.qdisc_stats.ecn_mark = q->cstats.ecn_mark; -+ st.qdisc_stats.new_flow_count = q->new_flow_count; -+ - list_for_each(pos, &q->new_flows) - st.qdisc_stats.new_flows_len++; - --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch b/src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch deleted file mode 100644 index 67bb750..0000000 --- a/src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch +++ /dev/null @@ -1,90 +0,0 @@ -From: Eric Dumazet <eric.dumazet(a)gmail.com> -Date: Sat, 12 May 2012 21:23:23 +0000 -Subject: [PATCH 6/7] codel: use u16 field instead of 31bits for rec_inv_sqrt - -commit 6ff272c9ad65eda219cd975b9da2dbc31cc812ee upstream. - -David pointed out gcc might generate poor code with 31bit fields. - -Using u16 is more than enough and permits a better code output. - -Also make the code intent more readable using constants, fixed point arithmetic -not being trivial for everybody. - -Suggested-by: David Miller <davem(a)davemloft.net> -Signed-off-by: Eric Dumazet <edumazet(a)google.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> ---- - include/net/codel.h | 25 +++++++++++++++---------- - 1 file changed, 15 insertions(+), 10 deletions(-) - -diff --git a/include/net/codel.h b/include/net/codel.h -index bd8747c..7546517 100644 ---- a/include/net/codel.h -+++ b/include/net/codel.h -@@ -133,13 +133,17 @@ struct codel_params { - struct codel_vars { - u32 count; - u32 lastcount; -- bool dropping:1; -- u32 rec_inv_sqrt:31; -+ bool dropping; -+ u16 rec_inv_sqrt; - codel_time_t first_above_time; - codel_time_t drop_next; - codel_time_t ldelay; - }; - -+#define REC_INV_SQRT_BITS (8 * sizeof(u16)) /* or sizeof_in_bits(rec_inv_sqrt) */ -+/* needed shift to get a Q0.32 number from rec_inv_sqrt */ -+#define REC_INV_SQRT_SHIFT (32 - REC_INV_SQRT_BITS) -+ - /** - * struct codel_stats - contains codel shared variables and stats - * @maxpacket: largest packet we've seen so far -@@ -173,17 +177,18 @@ static void codel_stats_init(struct codel_stats *stats) - * http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_me… - * new_invsqrt = (invsqrt / 2) * (3 - count * invsqrt^2) - * -- * Here, invsqrt is a fixed point number (< 1.0), 31bit mantissa) -+ * Here, invsqrt is a fixed point number (< 1.0), 32bit mantissa, aka Q0.32 - */ - static void codel_Newton_step(struct codel_vars *vars) - { -- u32 invsqrt = vars->rec_inv_sqrt; -- u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 31; -- u64 val = (3LL << 31) - ((u64)vars->count * invsqrt2); -+ u32 invsqrt = ((u32)vars->rec_inv_sqrt) << REC_INV_SQRT_SHIFT; -+ u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 32; -+ u64 val = (3LL << 32) - ((u64)vars->count * invsqrt2); - -- val = (val * invsqrt) >> 32; -+ val >>= 2; /* avoid overflow in following multiply */ -+ val = (val * invsqrt) >> (32 - 2 + 1); - -- vars->rec_inv_sqrt = val; -+ vars->rec_inv_sqrt = val >> REC_INV_SQRT_SHIFT; - } - - /* -@@ -195,7 +200,7 @@ static codel_time_t codel_control_law(codel_time_t t, - codel_time_t interval, - u32 rec_inv_sqrt) - { -- return t + reciprocal_divide(interval, rec_inv_sqrt << 1); -+ return t + reciprocal_divide(interval, rec_inv_sqrt << REC_INV_SQRT_SHIFT); - } - - -@@ -326,7 +331,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - codel_Newton_step(vars); - } else { - vars->count = 1; -- vars->rec_inv_sqrt = 0x7fffffff; -+ vars->rec_inv_sqrt = ~0U >> REC_INV_SQRT_SHIFT; - } - vars->lastcount = vars->count; - vars->drop_next = codel_control_law(now, params->interval, --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch b/src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch deleted file mode 100644 index 0195fdf..0000000 --- a/src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch +++ /dev/null @@ -1,140 +0,0 @@ -From: Eric Dumazet <edumazet(a)google.com> -Date: Wed, 16 May 2012 04:39:09 +0000 -Subject: [PATCH 7/7] fq_codel: should use qdisc backlog as threshold - -commit 865ec5523dadbedefbc5710a68969f686a28d928 upstream. - -codel_should_drop() logic allows a packet being not dropped if queue -size is under max packet size. - -In fq_codel, we have two possible backlogs : The qdisc global one, and -the flow local one. - -The meaningful one for codel_should_drop() should be the global backlog, -not the per flow one, so that thin flows can have a non zero drop/mark -probability. - -Signed-off-by: Eric Dumazet <edumazet(a)google.com> -Cc: Dave Taht <dave.taht(a)bufferbloat.net> -Cc: Kathleen Nichols <nichols(a)pollere.com> -Cc: Van Jacobson <van(a)pollere.net> -Signed-off-by: David S. Miller <davem(a)davemloft.net> ---- - include/net/codel.h | 15 +++++++-------- - net/sched/sch_codel.c | 4 ++-- - net/sched/sch_fq_codel.c | 5 +++-- - 3 files changed, 12 insertions(+), 12 deletions(-) - -diff --git a/include/net/codel.h b/include/net/codel.h -index 7546517..550debf 100644 ---- a/include/net/codel.h -+++ b/include/net/codel.h -@@ -205,7 +205,7 @@ static codel_time_t codel_control_law(codel_time_t t, - - - static bool codel_should_drop(const struct sk_buff *skb, -- unsigned int *backlog, -+ struct Qdisc *sch, - struct codel_vars *vars, - struct codel_params *params, - struct codel_stats *stats, -@@ -219,13 +219,13 @@ static bool codel_should_drop(const struct sk_buff *skb, - } - - vars->ldelay = now - codel_get_enqueue_time(skb); -- *backlog -= qdisc_pkt_len(skb); -+ sch->qstats.backlog -= qdisc_pkt_len(skb); - - if (unlikely(qdisc_pkt_len(skb) > stats->maxpacket)) - stats->maxpacket = qdisc_pkt_len(skb); - - if (codel_time_before(vars->ldelay, params->target) || -- *backlog <= stats->maxpacket) { -+ sch->qstats.backlog <= stats->maxpacket) { - /* went below - stay below for at least interval */ - vars->first_above_time = 0; - return false; -@@ -249,8 +249,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - struct codel_params *params, - struct codel_vars *vars, - struct codel_stats *stats, -- codel_skb_dequeue_t dequeue_func, -- u32 *backlog) -+ codel_skb_dequeue_t dequeue_func) - { - struct sk_buff *skb = dequeue_func(vars, sch); - codel_time_t now; -@@ -261,7 +260,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - return skb; - } - now = codel_get_time(); -- drop = codel_should_drop(skb, backlog, vars, params, stats, now); -+ drop = codel_should_drop(skb, sch, vars, params, stats, now); - if (vars->dropping) { - if (!drop) { - /* sojourn time below target - leave dropping state */ -@@ -292,7 +291,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - qdisc_drop(skb, sch); - stats->drop_count++; - skb = dequeue_func(vars, sch); -- if (!codel_should_drop(skb, backlog, -+ if (!codel_should_drop(skb, sch, - vars, params, stats, now)) { - /* leave dropping state */ - vars->dropping = false; -@@ -313,7 +312,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, - stats->drop_count++; - - skb = dequeue_func(vars, sch); -- drop = codel_should_drop(skb, backlog, vars, params, -+ drop = codel_should_drop(skb, sch, vars, params, - stats, now); - } - vars->dropping = true; -diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c -index 213ef60..2f9ab17 100644 ---- a/net/sched/sch_codel.c -+++ b/net/sched/sch_codel.c -@@ -77,8 +77,8 @@ static struct sk_buff *codel_qdisc_dequeue(struct Qdisc *sch) - struct codel_sched_data *q = qdisc_priv(sch); - struct sk_buff *skb; - -- skb = codel_dequeue(sch, &q->params, &q->vars, &q->stats, -- dequeue, &sch->qstats.backlog); -+ skb = codel_dequeue(sch, &q->params, &q->vars, &q->stats, dequeue); -+ - /* We cant call qdisc_tree_decrease_qlen() if our qlen is 0, - * or HTB crashes. Defer it for next round. - */ -diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c -index 337ff20..9fc1c62 100644 ---- a/net/sched/sch_fq_codel.c -+++ b/net/sched/sch_fq_codel.c -@@ -217,13 +217,14 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch) - */ - static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch) - { -+ struct fq_codel_sched_data *q = qdisc_priv(sch); - struct fq_codel_flow *flow; - struct sk_buff *skb = NULL; - - flow = container_of(vars, struct fq_codel_flow, cvars); - if (flow->head) { - skb = dequeue_head(flow); -- sch->qstats.backlog -= qdisc_pkt_len(skb); -+ q->backlogs[flow - q->flows] -= qdisc_pkt_len(skb); - sch->q.qlen--; - } - return skb; -@@ -256,7 +257,7 @@ begin: - prev_ecn_mark = q->cstats.ecn_mark; - - skb = codel_dequeue(sch, &q->cparams, &flow->cvars, &q->cstats, -- dequeue, &q->backlogs[flow - q->flows]); -+ dequeue); - - flow->dropped += q->cstats.drop_count - prev_drop_count; - flow->dropped += q->cstats.ecn_mark - prev_ecn_mark; --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch b/src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch deleted file mode 100644 index c88f253..0000000 --- a/src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch +++ /dev/null @@ -1,90 +0,0 @@ -From: Eric Dumazet <eric.dumazet(a)gmail.com> -Date: Mon, 28 Nov 2011 20:30:35 +0000 -Subject: [PATCH 2/3] flow_dissector: use a 64bit load/store -MIME-Version: 1.0 -Content-Type: text/plain; charset=UTF-8 -Content-Transfer-Encoding: 8bit - -commit 4d77d2b567ec66a443792d99e96ac760991d80d0 upstream. - -Le lundi 28 novembre 2011 à 19:06 -0500, David Miller a écrit : -> From: Dimitris Michailidis <dm(a)chelsio.com> -> Date: Mon, 28 Nov 2011 08:25:39 -0800 -> -> >> +bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys -> >> *flow) -> >> +{ -> >> + int poff, nhoff = skb_network_offset(skb); -> >> + u8 ip_proto; -> >> + u16 proto = skb->protocol; -> > -> > __be16 instead of u16 for proto? -> -> I'll take care of this when I apply these patches. - -( CC trimmed ) - -Thanks David ! - -Here is a small patch to use one 64bit load/store on x86_64 instead of -two 32bit load/stores. - -[PATCH net-next] flow_dissector: use a 64bit load/store - -gcc compiler is smart enough to use a single load/store if we -memcpy(dptr, sptr, 8) on x86_64, regardless of -CONFIG_CC_OPTIMIZE_FOR_SIZE - -In IP header, daddr immediately follows saddr, this wont change in the -future. We only need to make sure our flow_keys (src,dst) fields wont -break the rule. - -Signed-off-by: Eric Dumazet <eric.dumazet(a)gmail.com> -Signed-off-by: David S. Miller <davem(a)davemloft.net> - -diff --git a/include/net/flow_keys.h b/include/net/flow_keys.h -index e4cb285..80461c1 100644 ---- a/include/net/flow_keys.h -+++ b/include/net/flow_keys.h -@@ -2,6 +2,7 @@ - #define _NET_FLOW_KEYS_H - - struct flow_keys { -+ /* (src,dst) must be grouped, in the same way than in IP header */ - __be32 src; - __be32 dst; - union { -diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c -index f0516d9..0985b9b 100644 ---- a/net/core/flow_dissector.c -+++ b/net/core/flow_dissector.c -@@ -8,6 +8,16 @@ - #include <linux/ppp_defs.h> - #include <net/flow_keys.h> - -+/* copy saddr & daddr, possibly using 64bit load/store -+ * Equivalent to : flow->src = iph->saddr; -+ * flow->dst = iph->daddr; -+ */ -+static void iph_to_flow_copy_addrs(struct flow_keys *flow, const struct iphdr *iph) -+{ -+ BUILD_BUG_ON(offsetof(typeof(*flow), dst) != -+ offsetof(typeof(*flow), src) + sizeof(flow->src)); -+ memcpy(&flow->src, &iph->saddr, sizeof(flow->src) + sizeof(flow->dst)); -+} - - bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow) - { -@@ -31,8 +41,7 @@ ip: - ip_proto = 0; - else - ip_proto = iph->protocol; -- flow->src = iph->saddr; -- flow->dst = iph->daddr; -+ iph_to_flow_copy_addrs(flow, iph); - nhoff += iph->ihl * 4; - break; - } --- -1.7.10 - diff --git a/src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch b/src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch deleted file mode 100644 index ae4521d..0000000 --- a/src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch +++ /dev/null @@ -1,25 +0,0 @@ -From: Jesper Dangaard Brouer <hawk(a)comx.dk> -Date: Tue, 24 Jan 2012 16:03:33 -0500 -Subject: [PATCH 3/3] net: flow_dissector.c missing include linux/export.h - -commit c452ed70771cea3af73d21a5914989137fbd28b8 upstream. - -The file net/core/flow_dissector.c seems to be missing -including linux/export.h. - -Signed-off-by: Jesper Dangaard Brouer <hawk(a)comx.dk> -Signed-off-by: David S. Miller <davem(a)davemloft.net> - -diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c -index 0985b9b..a225089 100644 ---- a/net/core/flow_dissector.c -+++ b/net/core/flow_dissector.c -@@ -1,4 +1,5 @@ - #include <linux/skbuff.h> -+#include <linux/export.h> - #include <linux/ip.h> - #include <linux/ipv6.h> - #include <linux/if_vlan.h> --- -1.7.10 - hooks/post-receive -- IPFire 2.x development tree

1 0

[git.ipfire.org] IPFire 3.x development tree branch, master, updated. 76eb1f9c86ba8675af5484c0225d565f563dad9c
by git＠ipfire.org 20 Aug '12

20 Aug '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "IPFire 3.x development tree". The branch, master has been updated via 76eb1f9c86ba8675af5484c0225d565f563dad9c (commit) via 70cde3063f1f7a05a7d77b5814dc14ba3450a4a0 (commit) from 4c08cf3ead71e1fe5d8571c0c1e1fb40ccc03201 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 76eb1f9c86ba8675af5484c0225d565f563dad9c Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Mon Aug 20 12:49:09 2012 +0200 gnutls: Update to 3.1.0. commit 70cde3063f1f7a05a7d77b5814dc14ba3450a4a0 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Mon Aug 20 12:48:43 2012 +0200 nettle: Update to 2.5. ----------------------------------------------------------------------- Summary of changes: gnutls/gnutls.nm | 6 +++--- nettle/nettle.nm | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-) Difference in files: diff --git a/gnutls/gnutls.nm b/gnutls/gnutls.nm index fcd21a0..fa829f3 100644 --- a/gnutls/gnutls.nm +++ b/gnutls/gnutls.nm @@ -4,7 +4,7 @@ ############################################################################### name = gnutls -version = 3.0.15 +version = 3.1.0 release = 1 groups = System/Libraries @@ -29,7 +29,7 @@ build libgcrypt-devel libtasn1-devel lzo-devel - nettle-devel >= 2.4 + nettle-devel >= 2.5 perl p11-kit-devel >= 0.4 readline-devel @@ -54,7 +54,7 @@ end packages package %{name} requires - nettle >= 2.4 + nettle >= 2.5 p11-kit >= 0.4 end end diff --git a/nettle/nettle.nm b/nettle/nettle.nm index df0e73a..83d4acb 100644 --- a/nettle/nettle.nm +++ b/nettle/nettle.nm @@ -4,8 +4,8 @@ ############################################################################### name = nettle -version = 2.4 -release = 2 +version = 2.5 +release = 1 groups = System/Libraries url = http://www.lysator.liu.se/~nisse/nettle/ hooks/post-receive -- IPFire 3.x development tree

1 0

[git.ipfire.org] IPFire 3.x development tree branch, master, updated. 4c08cf3ead71e1fe5d8571c0c1e1fb40ccc03201
by git＠ipfire.org 18 Aug '12

18 Aug '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "IPFire 3.x development tree". The branch, master has been updated via 4c08cf3ead71e1fe5d8571c0c1e1fb40ccc03201 (commit) from 5425c9c9792c144a3f8a38df170303d0a0d41ee0 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 4c08cf3ead71e1fe5d8571c0c1e1fb40ccc03201 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sat Aug 18 14:29:11 2012 +0200 iproute2: Update to 3.5.1. ----------------------------------------------------------------------- Summary of changes: iproute2/iproute2.nm | 7 +++++-- 1 files changed, 5 insertions(+), 2 deletions(-) Difference in files: diff --git a/iproute2/iproute2.nm b/iproute2/iproute2.nm index 31dd9da..433e35f 100644 --- a/iproute2/iproute2.nm +++ b/iproute2/iproute2.nm @@ -4,7 +4,7 @@ ############################################################################### name = iproute2 -version = 3.4.0 +version = 3.5.1 release = 1 groups = Networking/Tools @@ -32,7 +32,7 @@ build end prepare_cmds - sed -e "s/_VERSION_/%{version}/" man/man8/ss.8 + sed -e "s/_VERSION_/%{version}/" -i man/man8/ss.8 end export RPM_OPT_FLAGS = %{CFLAGS} @@ -82,6 +82,9 @@ build # Install configuration. install -v -m 644 etc/iproute2/* %{BUILDROOT}%{sysconfdir}/iproute2 + + # bridge should be installed as 'br', following upstream + install -v -m 755 bridge/bridge %{BUILDROOT}%{sbindir}/br end # Keep libnetlink.a. hooks/post-receive -- IPFire 3.x development tree

1 0

[git.ipfire.org] IPFire 2.x development tree branch, kernel-update, updated. 3e78d712ccafa64c78439c02d2fae5a3615b0cd5
by git＠ipfire.org 14 Aug '12

14 Aug '12

1 0

[git.ipfire.org] IPFire 2.x development tree branch, kernel-update, updated. 97bb1c57aea0cfdb554a7bb96d1a121a868b6fa3
by git＠ipfire.org 13 Aug '12

13 Aug '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "IPFire 2.x development tree". The branch, kernel-update has been updated via 97bb1c57aea0cfdb554a7bb96d1a121a868b6fa3 (commit) from 21df33072c3592bc445760b7c72e0e81e0b74d78 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 97bb1c57aea0cfdb554a7bb96d1a121a868b6fa3 Author: Arne Fitzenreiter <Arne_F(a)ipfire.org> Date: Mon Aug 13 22:59:24 2012 +0200 kernel: add backported codel patchset. ----------------------------------------------------------------------- Summary of changes: .../kernel/kernel.config.armv5tel-ipfire-kirkwood | 2 + config/kernel/kernel.config.armv5tel-ipfire-omap | 2 + config/kernel/kernel.config.armv5tel-ipfire-rpi | 2 + config/kernel/kernel.config.i586-ipfire | 4 +- config/kernel/kernel.config.i586-ipfire-pae | 4 +- lfs/linux | 3 + .../0000-net-introduce-skb_flow_dissect.patch | 195 +++++ .../0001-codel-Controlled-Delay-AQM.patch | 771 ++++++++++++++++++ ...Newton-method-instead-of-sqrt-and-divides.patch | 189 +++++ .../0003-fq_codel-Fair-Queue-Codel-AQM.patch | 850 ++++++++++++++++++++ ...odel-Add-missing-include-linux-prefetch.h.patch | 37 + .../0005-net-codel-fix-build-errors.patch | 55 ++ ...16-field-instead-of-31bits-for-rec_inv_sq.patch | 90 ++ ...del-should-use-qdisc-backlog-as-threshold.patch | 140 ++++ ...008-flow_dissector-use-a-64bit-load-store.patch | 90 ++ ...issector.c-missing-include-linux-export.h.patch | 25 + 16 files changed, 2457 insertions(+), 2 deletions(-) create mode 100644 src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch create mode 100644 src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch create mode 100644 src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch create mode 100644 src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch create mode 100644 src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch create mode 100644 src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch create mode 100644 src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch create mode 100644 src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch create mode 100644 src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch create mode 100644 src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch Difference in files: diff --git a/config/kernel/kernel.config.armv5tel-ipfire-kirkwood b/config/kernel/kernel.config.armv5tel-ipfire-kirkwood index 133ac1a..67992fd 100644 --- a/config/kernel/kernel.config.armv5tel-ipfire-kirkwood +++ b/config/kernel/kernel.config.armv5tel-ipfire-kirkwood @@ -832,6 +832,8 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m +CONFIG_NET_SCH_CODEL=m +CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.armv5tel-ipfire-omap b/config/kernel/kernel.config.armv5tel-ipfire-omap index 08405e2..17d5109 100644 --- a/config/kernel/kernel.config.armv5tel-ipfire-omap +++ b/config/kernel/kernel.config.armv5tel-ipfire-omap @@ -924,6 +924,8 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m +CONFIG_NET_SCH_CODEL=m +CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.armv5tel-ipfire-rpi b/config/kernel/kernel.config.armv5tel-ipfire-rpi index 4d434e0..414dd8f 100644 --- a/config/kernel/kernel.config.armv5tel-ipfire-rpi +++ b/config/kernel/kernel.config.armv5tel-ipfire-rpi @@ -798,6 +798,8 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m +CONFIG_NET_SCH_CODEL=m +CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.i586-ipfire b/config/kernel/kernel.config.i586-ipfire index 4295eb4..1581ea0 100644 --- a/config/kernel/kernel.config.i586-ipfire +++ b/config/kernel/kernel.config.i586-ipfire @@ -1,6 +1,6 @@ # # Automatically generated file; DO NOT EDIT. -# Linux/i386 3.2.20 Kernel Configuration +# Linux/i386 3.2.27 Kernel Configuration # # CONFIG_64BIT is not set CONFIG_X86_32=y @@ -973,6 +973,8 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m +CONFIG_NET_SCH_CODEL=m +CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/config/kernel/kernel.config.i586-ipfire-pae b/config/kernel/kernel.config.i586-ipfire-pae index 7186e73..4157b09 100644 --- a/config/kernel/kernel.config.i586-ipfire-pae +++ b/config/kernel/kernel.config.i586-ipfire-pae @@ -1,6 +1,6 @@ # # Automatically generated file; DO NOT EDIT. -# Linux/i386 2.6.42.20 Kernel Configuration +# Linux/i386 3.2.27 Kernel Configuration # # CONFIG_64BIT is not set CONFIG_X86_32=y @@ -996,6 +996,8 @@ CONFIG_NET_SCH_DRR=m CONFIG_NET_SCH_MQPRIO=m CONFIG_NET_SCH_CHOKE=m CONFIG_NET_SCH_QFQ=m +CONFIG_NET_SCH_CODEL=m +CONFIG_NET_SCH_FQ_CODEL=m CONFIG_NET_SCH_INGRESS=m # diff --git a/lfs/linux b/lfs/linux index 9f994f8..2c92809 100644 --- a/lfs/linux +++ b/lfs/linux @@ -111,6 +111,9 @@ $(TARGET) : $(patsubst %,$(DIR_DL)/%,$(objects)) ln -svf linux-$(VER) $(DIR_SRC)/linux + # Apply backported codel patches + cd $(DIR_APP) && cat $(DIR_SRC)/src/patches/linux-3.2-codel/*.patch | patch -Np1 + # Linux Intermediate Queueing Device cd $(DIR_APP) && patch -Np1 < $(DIR_SRC)/src/patches/linux-3.2-imq.patch diff --git a/src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch b/src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch new file mode 100644 index 0000000..a749507 --- /dev/null +++ b/src/patches/linux-3.2-codel/0000-net-introduce-skb_flow_dissect.patch @@ -0,0 +1,195 @@ +From: Eric Dumazet <eric.dumazet(a)gmail.com> +Date: Mon, 28 Nov 2011 05:22:18 +0000 +Subject: [PATCH 1/3] net: introduce skb_flow_dissect() + +commit 0744dd00c1b1be99a25b62b1b48df440e82e57e0 upstream. + +We use at least two flow dissectors in network stack, with known +limitations and code duplication. + +Introduce skb_flow_dissect() to factorize this, highly inspired from +existing dissector from __skb_get_rxhash() + +Note : We extensively use skb_header_pointer(), this permits us to not +touch skb at all. + +Signed-off-by: Eric Dumazet <eric.dumazet(a)gmail.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> + +diff --git a/include/net/flow_keys.h b/include/net/flow_keys.h +new file mode 100644 +index 0000000..e4cb285 +--- /dev/null ++++ b/include/net/flow_keys.h +@@ -0,0 +1,15 @@ ++#ifndef _NET_FLOW_KEYS_H ++#define _NET_FLOW_KEYS_H ++ ++struct flow_keys { ++ __be32 src; ++ __be32 dst; ++ union { ++ __be32 ports; ++ __be16 port16[2]; ++ }; ++ u8 ip_proto; ++}; ++ ++extern bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow); ++#endif +diff --git a/net/core/Makefile b/net/core/Makefile +index 3606d40..c4ecc86 100644 +--- a/net/core/Makefile ++++ b/net/core/Makefile +@@ -3,7 +3,7 @@ + # + + obj-y := sock.o request_sock.o skbuff.o iovec.o datagram.o stream.o scm.o \ +- gen_stats.o gen_estimator.o net_namespace.o secure_seq.o ++ gen_stats.o gen_estimator.o net_namespace.o secure_seq.o flow_dissector.o + + obj-$(CONFIG_SYSCTL) += sysctl_net_core.o + +diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c +new file mode 100644 +index 0000000..f0516d9 +--- /dev/null ++++ b/net/core/flow_dissector.c +@@ -0,0 +1,134 @@ ++#include <linux/skbuff.h> ++#include <linux/ip.h> ++#include <linux/ipv6.h> ++#include <linux/if_vlan.h> ++#include <net/ip.h> ++#include <linux/if_tunnel.h> ++#include <linux/if_pppox.h> ++#include <linux/ppp_defs.h> ++#include <net/flow_keys.h> ++ ++ ++bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow) ++{ ++ int poff, nhoff = skb_network_offset(skb); ++ u8 ip_proto; ++ __be16 proto = skb->protocol; ++ ++ memset(flow, 0, sizeof(*flow)); ++ ++again: ++ switch (proto) { ++ case __constant_htons(ETH_P_IP): { ++ const struct iphdr *iph; ++ struct iphdr _iph; ++ip: ++ iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph); ++ if (!iph) ++ return false; ++ ++ if (ip_is_fragment(iph)) ++ ip_proto = 0; ++ else ++ ip_proto = iph->protocol; ++ flow->src = iph->saddr; ++ flow->dst = iph->daddr; ++ nhoff += iph->ihl * 4; ++ break; ++ } ++ case __constant_htons(ETH_P_IPV6): { ++ const struct ipv6hdr *iph; ++ struct ipv6hdr _iph; ++ipv6: ++ iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph); ++ if (!iph) ++ return false; ++ ++ ip_proto = iph->nexthdr; ++ flow->src = iph->saddr.s6_addr32[3]; ++ flow->dst = iph->daddr.s6_addr32[3]; ++ nhoff += sizeof(struct ipv6hdr); ++ break; ++ } ++ case __constant_htons(ETH_P_8021Q): { ++ const struct vlan_hdr *vlan; ++ struct vlan_hdr _vlan; ++ ++ vlan = skb_header_pointer(skb, nhoff, sizeof(_vlan), &_vlan); ++ if (!vlan) ++ return false; ++ ++ proto = vlan->h_vlan_encapsulated_proto; ++ nhoff += sizeof(*vlan); ++ goto again; ++ } ++ case __constant_htons(ETH_P_PPP_SES): { ++ struct { ++ struct pppoe_hdr hdr; ++ __be16 proto; ++ } *hdr, _hdr; ++ hdr = skb_header_pointer(skb, nhoff, sizeof(_hdr), &_hdr); ++ if (!hdr) ++ return false; ++ proto = hdr->proto; ++ nhoff += PPPOE_SES_HLEN; ++ switch (proto) { ++ case __constant_htons(PPP_IP): ++ goto ip; ++ case __constant_htons(PPP_IPV6): ++ goto ipv6; ++ default: ++ return false; ++ } ++ } ++ default: ++ return false; ++ } ++ ++ switch (ip_proto) { ++ case IPPROTO_GRE: { ++ struct gre_hdr { ++ __be16 flags; ++ __be16 proto; ++ } *hdr, _hdr; ++ ++ hdr = skb_header_pointer(skb, nhoff, sizeof(_hdr), &_hdr); ++ if (!hdr) ++ return false; ++ /* ++ * Only look inside GRE if version zero and no ++ * routing ++ */ ++ if (!(hdr->flags & (GRE_VERSION|GRE_ROUTING))) { ++ proto = hdr->proto; ++ nhoff += 4; ++ if (hdr->flags & GRE_CSUM) ++ nhoff += 4; ++ if (hdr->flags & GRE_KEY) ++ nhoff += 4; ++ if (hdr->flags & GRE_SEQ) ++ nhoff += 4; ++ goto again; ++ } ++ break; ++ } ++ case IPPROTO_IPIP: ++ goto again; ++ default: ++ break; ++ } ++ ++ flow->ip_proto = ip_proto; ++ poff = proto_ports_offset(ip_proto); ++ if (poff >= 0) { ++ __be32 *ports, _ports; ++ ++ nhoff += poff; ++ ports = skb_header_pointer(skb, nhoff, sizeof(_ports), &_ports); ++ if (ports) ++ flow->ports = *ports; ++ } ++ ++ return true; ++} ++EXPORT_SYMBOL(skb_flow_dissect); +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch b/src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch new file mode 100644 index 0000000..0d8d18b --- /dev/null +++ b/src/patches/linux-3.2-codel/0001-codel-Controlled-Delay-AQM.patch @@ -0,0 +1,771 @@ +From: Eric Dumazet <edumazet(a)google.com> +Date: Thu, 10 May 2012 07:51:25 +0000 +Subject: [PATCH 1/7] codel: Controlled Delay AQM + +commit 76e3cc126bb223013a6b9a0e2a51238d1ef2e409 upstream. + +An implementation of CoDel AQM, from Kathleen Nichols and Van Jacobson. + +http://queue.acm.org/detail.cfm?id=2209336 + +This AQM main input is no longer queue size in bytes or packets, but the +delay packets stay in (FIFO) queue. + +As we don't have infinite memory, we still can drop packets in enqueue() +in case of massive load, but mean of CoDel is to drop packets in +dequeue(), using a control law based on two simple parameters : + +target : target sojourn time (default 5ms) +interval : width of moving time window (default 100ms) + +Based on initial work from Dave Taht. + +Refactored to help future codel inclusion as a plugin for other linux +qdisc (FQ_CODEL, ...), like RED. + +include/net/codel.h contains codel algorithm as close as possible than +Kathleen reference. + +net/sched/sch_codel.c contains the linux qdisc specific glue. + +Separate structures permit a memory efficient implementation of fq_codel +(to be sent as a separate work) : Each flow has its own struct +codel_vars. + +timestamps are taken at enqueue() time with 1024 ns precision, allowing +a range of 2199 seconds in queue, and 100Gb links support. iproute2 uses +usec as base unit. + +Selected packets are dropped, unless ECN is enabled and packets can get +ECN mark instead. + +Tested from 2Mb to 10Gb speeds with no particular problems, on ixgbe and +tg3 drivers (BQL enabled). + +Usage: tc qdisc ... codel [ limit PACKETS ] [ target TIME ] + [ interval TIME ] [ ecn ] + +qdisc codel 10: parent 1:1 limit 2000p target 3.0ms interval 60.0ms ecn + Sent 13347099587 bytes 8815805 pkt (dropped 0, overlimits 0 requeues 0) + rate 202365Kbit 16708pps backlog 113550b 75p requeues 0 + count 116 lastcount 98 ldelay 4.3ms dropping drop_next 816us + maxpacket 1514 ecn_mark 84399 drop_overlimit 0 + +CoDel must be seen as a base module, and should be used keeping in mind +there is still a FIFO queue. So a typical setup will probably need a +hierarchy of several qdiscs and packet classifiers to be able to meet +whatever constraints a user might have. + +One possible example would be to use fq_codel, which combines Fair +Queueing and CoDel, in replacement of sfq / sfq_red. + +Signed-off-by: Eric Dumazet <edumazet(a)google.com> +Signed-off-by: Dave Taht <dave.taht(a)bufferbloat.net> +Cc: Kathleen Nichols <nichols(a)pollere.com> +Cc: Van Jacobson <van(a)pollere.net> +Cc: Tom Herbert <therbert(a)google.com> +Cc: Matt Mathis <mattmathis(a)google.com> +Cc: Yuchung Cheng <ycheng(a)google.com> +Cc: Stephen Hemminger <shemminger(a)vyatta.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> +--- + include/linux/pkt_sched.h | 26 ++++ + include/net/codel.h | 332 +++++++++++++++++++++++++++++++++++++++++++++ + net/sched/Kconfig | 11 ++ + net/sched/Makefile | 1 + + net/sched/sch_codel.c | 275 +++++++++++++++++++++++++++++++++++++ + 5 files changed, 645 insertions(+) + create mode 100644 include/net/codel.h + create mode 100644 net/sched/sch_codel.c + +diff --git a/include/linux/pkt_sched.h b/include/linux/pkt_sched.h +index ffe975c..cde56c2 100644 +--- a/include/linux/pkt_sched.h ++++ b/include/linux/pkt_sched.h +@@ -655,4 +655,30 @@ struct tc_qfq_stats { + __u32 lmax; + }; + ++/* CODEL */ ++ ++enum { ++ TCA_CODEL_UNSPEC, ++ TCA_CODEL_TARGET, ++ TCA_CODEL_LIMIT, ++ TCA_CODEL_INTERVAL, ++ TCA_CODEL_ECN, ++ __TCA_CODEL_MAX ++}; ++ ++#define TCA_CODEL_MAX (__TCA_CODEL_MAX - 1) ++ ++struct tc_codel_xstats { ++ __u32 maxpacket; /* largest packet we've seen so far */ ++ __u32 count; /* how many drops we've done since the last time we ++ * entered dropping state ++ */ ++ __u32 lastcount; /* count at entry to dropping state */ ++ __u32 ldelay; /* in-queue delay seen by most recently dequeued packet */ ++ __s32 drop_next; /* time to drop next packet */ ++ __u32 drop_overlimit; /* number of time max qdisc packet limit was hit */ ++ __u32 ecn_mark; /* number of packets we ECN marked instead of dropped */ ++ __u32 dropping; /* are we in dropping state ? */ ++}; ++ + #endif +diff --git a/include/net/codel.h b/include/net/codel.h +new file mode 100644 +index 0000000..bce2cef +--- /dev/null ++++ b/include/net/codel.h +@@ -0,0 +1,332 @@ ++#ifndef __NET_SCHED_CODEL_H ++#define __NET_SCHED_CODEL_H ++ ++/* ++ * Codel - The Controlled-Delay Active Queue Management algorithm ++ * ++ * Copyright (C) 2011-2012 Kathleen Nichols <nichols(a)pollere.com> ++ * Copyright (C) 2011-2012 Van Jacobson <van(a)pollere.net> ++ * Copyright (C) 2012 Michael D. Taht <dave.taht(a)bufferbloat.net> ++ * Copyright (C) 2012 Eric Dumazet <edumazet(a)google.com> ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions, and the following disclaimer, ++ * without modification. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The names of the authors may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * Alternatively, provided that this notice is retained in full, this ++ * software may be distributed under the terms of the GNU General ++ * Public License ("GPL") version 2, in which case the provisions of the ++ * GPL apply INSTEAD OF those given above. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH ++ * DAMAGE. ++ * ++ */ ++ ++#include <linux/types.h> ++#include <linux/ktime.h> ++#include <linux/skbuff.h> ++#include <net/pkt_sched.h> ++#include <net/inet_ecn.h> ++ ++/* Controlling Queue Delay (CoDel) algorithm ++ * ========================================= ++ * Source : Kathleen Nichols and Van Jacobson ++ * http://queue.acm.org/detail.cfm?id=2209336 ++ * ++ * Implemented on linux by Dave Taht and Eric Dumazet ++ */ ++ ++ ++/* CoDel uses a 1024 nsec clock, encoded in u32 ++ * This gives a range of 2199 seconds, because of signed compares ++ */ ++typedef u32 codel_time_t; ++typedef s32 codel_tdiff_t; ++#define CODEL_SHIFT 10 ++#define MS2TIME(a) ((a * NSEC_PER_MSEC) >> CODEL_SHIFT) ++ ++static inline codel_time_t codel_get_time(void) ++{ ++ u64 ns = ktime_to_ns(ktime_get()); ++ ++ return ns >> CODEL_SHIFT; ++} ++ ++#define codel_time_after(a, b) ((s32)(a) - (s32)(b) > 0) ++#define codel_time_after_eq(a, b) ((s32)(a) - (s32)(b) >= 0) ++#define codel_time_before(a, b) ((s32)(a) - (s32)(b) < 0) ++#define codel_time_before_eq(a, b) ((s32)(a) - (s32)(b) <= 0) ++ ++/* Qdiscs using codel plugin must use codel_skb_cb in their own cb[] */ ++struct codel_skb_cb { ++ codel_time_t enqueue_time; ++}; ++ ++static struct codel_skb_cb *get_codel_cb(const struct sk_buff *skb) ++{ ++ qdisc_cb_private_validate(skb, sizeof(struct codel_skb_cb)); ++ return (struct codel_skb_cb *)qdisc_skb_cb(skb)->data; ++} ++ ++static codel_time_t codel_get_enqueue_time(const struct sk_buff *skb) ++{ ++ return get_codel_cb(skb)->enqueue_time; ++} ++ ++static void codel_set_enqueue_time(struct sk_buff *skb) ++{ ++ get_codel_cb(skb)->enqueue_time = codel_get_time(); ++} ++ ++static inline u32 codel_time_to_us(codel_time_t val) ++{ ++ u64 valns = ((u64)val << CODEL_SHIFT); ++ ++ do_div(valns, NSEC_PER_USEC); ++ return (u32)valns; ++} ++ ++/** ++ * struct codel_params - contains codel parameters ++ * @target: target queue size (in time units) ++ * @interval: width of moving time window ++ * @ecn: is Explicit Congestion Notification enabled ++ */ ++struct codel_params { ++ codel_time_t target; ++ codel_time_t interval; ++ bool ecn; ++}; ++ ++/** ++ * struct codel_vars - contains codel variables ++ * @count: how many drops we've done since the last time we ++ * entered dropping state ++ * @lastcount: count at entry to dropping state ++ * @dropping: set to true if in dropping state ++ * @first_above_time: when we went (or will go) continuously above target ++ * for interval ++ * @drop_next: time to drop next packet, or when we dropped last ++ * @ldelay: sojourn time of last dequeued packet ++ */ ++struct codel_vars { ++ u32 count; ++ u32 lastcount; ++ bool dropping; ++ codel_time_t first_above_time; ++ codel_time_t drop_next; ++ codel_time_t ldelay; ++}; ++ ++/** ++ * struct codel_stats - contains codel shared variables and stats ++ * @maxpacket: largest packet we've seen so far ++ * @drop_count: temp count of dropped packets in dequeue() ++ * ecn_mark: number of packets we ECN marked instead of dropping ++ */ ++struct codel_stats { ++ u32 maxpacket; ++ u32 drop_count; ++ u32 ecn_mark; ++}; ++ ++static void codel_params_init(struct codel_params *params) ++{ ++ params->interval = MS2TIME(100); ++ params->target = MS2TIME(5); ++ params->ecn = false; ++} ++ ++static void codel_vars_init(struct codel_vars *vars) ++{ ++ vars->drop_next = 0; ++ vars->first_above_time = 0; ++ vars->dropping = false; /* exit dropping state */ ++ vars->count = 0; ++ vars->lastcount = 0; ++} ++ ++static void codel_stats_init(struct codel_stats *stats) ++{ ++ stats->maxpacket = 256; ++} ++ ++/* return interval/sqrt(x) with good precision ++ * relies on int_sqrt(unsigned long x) kernel implementation ++ */ ++static u32 codel_inv_sqrt(u32 _interval, u32 _x) ++{ ++ u64 interval = _interval; ++ unsigned long x = _x; ++ ++ /* Scale operands for max precision */ ++ ++#if BITS_PER_LONG == 64 ++ x <<= 32; /* On 64bit arches, we can prescale x by 32bits */ ++ interval <<= 16; ++#endif ++ ++ while (x < (1UL << (BITS_PER_LONG - 2))) { ++ x <<= 2; ++ interval <<= 1; ++ } ++ do_div(interval, int_sqrt(x)); ++ return (u32)interval; ++} ++ ++static codel_time_t codel_control_law(codel_time_t t, ++ codel_time_t interval, ++ u32 count) ++{ ++ return t + codel_inv_sqrt(interval, count); ++} ++ ++ ++static bool codel_should_drop(struct sk_buff *skb, ++ unsigned int *backlog, ++ struct codel_vars *vars, ++ struct codel_params *params, ++ struct codel_stats *stats, ++ codel_time_t now) ++{ ++ bool ok_to_drop; ++ ++ if (!skb) { ++ vars->first_above_time = 0; ++ return false; ++ } ++ ++ vars->ldelay = now - codel_get_enqueue_time(skb); ++ *backlog -= qdisc_pkt_len(skb); ++ ++ if (unlikely(qdisc_pkt_len(skb) > stats->maxpacket)) ++ stats->maxpacket = qdisc_pkt_len(skb); ++ ++ if (codel_time_before(vars->ldelay, params->target) || ++ *backlog <= stats->maxpacket) { ++ /* went below - stay below for at least interval */ ++ vars->first_above_time = 0; ++ return false; ++ } ++ ok_to_drop = false; ++ if (vars->first_above_time == 0) { ++ /* just went above from below. If we stay above ++ * for at least interval we'll say it's ok to drop ++ */ ++ vars->first_above_time = now + params->interval; ++ } else if (codel_time_after(now, vars->first_above_time)) { ++ ok_to_drop = true; ++ } ++ return ok_to_drop; ++} ++ ++typedef struct sk_buff * (*codel_skb_dequeue_t)(struct codel_vars *vars, ++ struct Qdisc *sch); ++ ++static struct sk_buff *codel_dequeue(struct Qdisc *sch, ++ struct codel_params *params, ++ struct codel_vars *vars, ++ struct codel_stats *stats, ++ codel_skb_dequeue_t dequeue_func, ++ u32 *backlog) ++{ ++ struct sk_buff *skb = dequeue_func(vars, sch); ++ codel_time_t now; ++ bool drop; ++ ++ if (!skb) { ++ vars->dropping = false; ++ return skb; ++ } ++ now = codel_get_time(); ++ drop = codel_should_drop(skb, backlog, vars, params, stats, now); ++ if (vars->dropping) { ++ if (!drop) { ++ /* sojourn time below target - leave dropping state */ ++ vars->dropping = false; ++ } else if (codel_time_after_eq(now, vars->drop_next)) { ++ /* It's time for the next drop. Drop the current ++ * packet and dequeue the next. The dequeue might ++ * take us out of dropping state. ++ * If not, schedule the next drop. ++ * A large backlog might result in drop rates so high ++ * that the next drop should happen now, ++ * hence the while loop. ++ */ ++ while (vars->dropping && ++ codel_time_after_eq(now, vars->drop_next)) { ++ if (++vars->count == 0) /* avoid zero divides */ ++ vars->count = ~0U; ++ if (params->ecn && INET_ECN_set_ce(skb)) { ++ stats->ecn_mark++; ++ vars->drop_next = ++ codel_control_law(vars->drop_next, ++ params->interval, ++ vars->count); ++ goto end; ++ } ++ qdisc_drop(skb, sch); ++ stats->drop_count++; ++ skb = dequeue_func(vars, sch); ++ if (!codel_should_drop(skb, backlog, ++ vars, params, stats, now)) { ++ /* leave dropping state */ ++ vars->dropping = false; ++ } else { ++ /* and schedule the next drop */ ++ vars->drop_next = ++ codel_control_law(vars->drop_next, ++ params->interval, ++ vars->count); ++ } ++ } ++ } ++ } else if (drop) { ++ if (params->ecn && INET_ECN_set_ce(skb)) { ++ stats->ecn_mark++; ++ } else { ++ qdisc_drop(skb, sch); ++ stats->drop_count++; ++ ++ skb = dequeue_func(vars, sch); ++ drop = codel_should_drop(skb, backlog, vars, params, ++ stats, now); ++ } ++ vars->dropping = true; ++ /* if min went above target close to when we last went below it ++ * assume that the drop rate that controlled the queue on the ++ * last cycle is a good starting point to control it now. ++ */ ++ if (codel_time_before(now - vars->drop_next, ++ 16 * params->interval)) { ++ vars->count = (vars->count - vars->lastcount) | 1; ++ } else { ++ vars->count = 1; ++ } ++ vars->lastcount = vars->count; ++ vars->drop_next = codel_control_law(now, params->interval, ++ vars->count); ++ } ++end: ++ return skb; ++} ++#endif +diff --git a/net/sched/Kconfig b/net/sched/Kconfig +index 75b58f8..fadd252 100644 +--- a/net/sched/Kconfig ++++ b/net/sched/Kconfig +@@ -250,6 +250,17 @@ config NET_SCH_QFQ + + If unsure, say N. + ++config NET_SCH_CODEL ++ tristate "Controlled Delay AQM (CODEL)" ++ help ++ Say Y here if you want to use the Controlled Delay (CODEL) ++ packet scheduling algorithm. ++ ++ To compile this driver as a module, choose M here: the module ++ will be called sch_codel. ++ ++ If unsure, say N. ++ + config NET_SCH_INGRESS + tristate "Ingress Qdisc" + depends on NET_CLS_ACT +diff --git a/net/sched/Makefile b/net/sched/Makefile +index 8cdf4e2..30fab03 100644 +--- a/net/sched/Makefile ++++ b/net/sched/Makefile +@@ -37,6 +37,7 @@ obj-$(CONFIG_NET_SCH_PLUG) += sch_plug.o + obj-$(CONFIG_NET_SCH_MQPRIO) += sch_mqprio.o + obj-$(CONFIG_NET_SCH_CHOKE) += sch_choke.o + obj-$(CONFIG_NET_SCH_QFQ) += sch_qfq.o ++obj-$(CONFIG_NET_SCH_CODEL) += sch_codel.o + + obj-$(CONFIG_NET_CLS_U32) += cls_u32.o + obj-$(CONFIG_NET_CLS_ROUTE4) += cls_route.o +diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c +new file mode 100644 +index 0000000..b4a1a81 +--- /dev/null ++++ b/net/sched/sch_codel.c +@@ -0,0 +1,275 @@ ++/* ++ * Codel - The Controlled-Delay Active Queue Management algorithm ++ * ++ * Copyright (C) 2011-2012 Kathleen Nichols <nichols(a)pollere.com> ++ * Copyright (C) 2011-2012 Van Jacobson <van(a)pollere.net> ++ * ++ * Implemented on linux by : ++ * Copyright (C) 2012 Michael D. Taht <dave.taht(a)bufferbloat.net> ++ * Copyright (C) 2012 Eric Dumazet <edumazet(a)google.com> ++ * ++ * Redistribution and use in source and binary forms, with or without ++ * modification, are permitted provided that the following conditions ++ * are met: ++ * 1. Redistributions of source code must retain the above copyright ++ * notice, this list of conditions, and the following disclaimer, ++ * without modification. ++ * 2. Redistributions in binary form must reproduce the above copyright ++ * notice, this list of conditions and the following disclaimer in the ++ * documentation and/or other materials provided with the distribution. ++ * 3. The names of the authors may not be used to endorse or promote products ++ * derived from this software without specific prior written permission. ++ * ++ * Alternatively, provided that this notice is retained in full, this ++ * software may be distributed under the terms of the GNU General ++ * Public License ("GPL") version 2, in which case the provisions of the ++ * GPL apply INSTEAD OF those given above. ++ * ++ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ++ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ++ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR ++ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT ++ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, ++ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT ++ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, ++ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY ++ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT ++ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE ++ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH ++ * DAMAGE. ++ * ++ */ ++ ++#include <linux/module.h> ++#include <linux/slab.h> ++#include <linux/types.h> ++#include <linux/kernel.h> ++#include <linux/errno.h> ++#include <linux/skbuff.h> ++#include <net/pkt_sched.h> ++#include <net/codel.h> ++ ++ ++#define DEFAULT_CODEL_LIMIT 1000 ++ ++struct codel_sched_data { ++ struct codel_params params; ++ struct codel_vars vars; ++ struct codel_stats stats; ++ u32 drop_overlimit; ++}; ++ ++/* This is the specific function called from codel_dequeue() ++ * to dequeue a packet from queue. Note: backlog is handled in ++ * codel, we dont need to reduce it here. ++ */ ++static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch) ++{ ++ struct sk_buff *skb = __skb_dequeue(&sch->q); ++ ++ prefetch(&skb->end); /* we'll need skb_shinfo() */ ++ return skb; ++} ++ ++static struct sk_buff *codel_qdisc_dequeue(struct Qdisc *sch) ++{ ++ struct codel_sched_data *q = qdisc_priv(sch); ++ struct sk_buff *skb; ++ ++ skb = codel_dequeue(sch, &q->params, &q->vars, &q->stats, ++ dequeue, &sch->qstats.backlog); ++ /* We cant call qdisc_tree_decrease_qlen() if our qlen is 0, ++ * or HTB crashes. Defer it for next round. ++ */ ++ if (q->stats.drop_count && sch->q.qlen) { ++ qdisc_tree_decrease_qlen(sch, q->stats.drop_count); ++ q->stats.drop_count = 0; ++ } ++ if (skb) ++ qdisc_bstats_update(sch, skb); ++ return skb; ++} ++ ++static int codel_qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch) ++{ ++ struct codel_sched_data *q; ++ ++ if (likely(qdisc_qlen(sch) < sch->limit)) { ++ codel_set_enqueue_time(skb); ++ return qdisc_enqueue_tail(skb, sch); ++ } ++ q = qdisc_priv(sch); ++ q->drop_overlimit++; ++ return qdisc_drop(skb, sch); ++} ++ ++static const struct nla_policy codel_policy[TCA_CODEL_MAX + 1] = { ++ [TCA_CODEL_TARGET] = { .type = NLA_U32 }, ++ [TCA_CODEL_LIMIT] = { .type = NLA_U32 }, ++ [TCA_CODEL_INTERVAL] = { .type = NLA_U32 }, ++ [TCA_CODEL_ECN] = { .type = NLA_U32 }, ++}; ++ ++static int codel_change(struct Qdisc *sch, struct nlattr *opt) ++{ ++ struct codel_sched_data *q = qdisc_priv(sch); ++ struct nlattr *tb[TCA_CODEL_MAX + 1]; ++ unsigned int qlen; ++ int err; ++ ++ if (!opt) ++ return -EINVAL; ++ ++ err = nla_parse_nested(tb, TCA_CODEL_MAX, opt, codel_policy); ++ if (err < 0) ++ return err; ++ ++ sch_tree_lock(sch); ++ ++ if (tb[TCA_CODEL_TARGET]) { ++ u32 target = nla_get_u32(tb[TCA_CODEL_TARGET]); ++ ++ q->params.target = ((u64)target * NSEC_PER_USEC) >> CODEL_SHIFT; ++ } ++ ++ if (tb[TCA_CODEL_INTERVAL]) { ++ u32 interval = nla_get_u32(tb[TCA_CODEL_INTERVAL]); ++ ++ q->params.interval = ((u64)interval * NSEC_PER_USEC) >> CODEL_SHIFT; ++ } ++ ++ if (tb[TCA_CODEL_LIMIT]) ++ sch->limit = nla_get_u32(tb[TCA_CODEL_LIMIT]); ++ ++ if (tb[TCA_CODEL_ECN]) ++ q->params.ecn = !!nla_get_u32(tb[TCA_CODEL_ECN]); ++ ++ qlen = sch->q.qlen; ++ while (sch->q.qlen > sch->limit) { ++ struct sk_buff *skb = __skb_dequeue(&sch->q); ++ ++ sch->qstats.backlog -= qdisc_pkt_len(skb); ++ qdisc_drop(skb, sch); ++ } ++ qdisc_tree_decrease_qlen(sch, qlen - sch->q.qlen); ++ ++ sch_tree_unlock(sch); ++ return 0; ++} ++ ++static int codel_init(struct Qdisc *sch, struct nlattr *opt) ++{ ++ struct codel_sched_data *q = qdisc_priv(sch); ++ ++ sch->limit = DEFAULT_CODEL_LIMIT; ++ ++ codel_params_init(&q->params); ++ codel_vars_init(&q->vars); ++ codel_stats_init(&q->stats); ++ ++ if (opt) { ++ int err = codel_change(sch, opt); ++ ++ if (err) ++ return err; ++ } ++ ++ if (sch->limit >= 1) ++ sch->flags |= TCQ_F_CAN_BYPASS; ++ else ++ sch->flags &= ~TCQ_F_CAN_BYPASS; ++ ++ return 0; ++} ++ ++static int codel_dump(struct Qdisc *sch, struct sk_buff *skb) ++{ ++ struct codel_sched_data *q = qdisc_priv(sch); ++ struct nlattr *opts; ++ ++ opts = nla_nest_start(skb, TCA_OPTIONS); ++ if (opts == NULL) ++ goto nla_put_failure; ++ ++ if (nla_put_u32(skb, TCA_CODEL_TARGET, ++ codel_time_to_us(q->params.target)) || ++ nla_put_u32(skb, TCA_CODEL_LIMIT, ++ sch->limit) || ++ nla_put_u32(skb, TCA_CODEL_INTERVAL, ++ codel_time_to_us(q->params.interval)) || ++ nla_put_u32(skb, TCA_CODEL_ECN, ++ q->params.ecn)) ++ goto nla_put_failure; ++ ++ return nla_nest_end(skb, opts); ++ ++nla_put_failure: ++ nla_nest_cancel(skb, opts); ++ return -1; ++} ++ ++static int codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d) ++{ ++ const struct codel_sched_data *q = qdisc_priv(sch); ++ struct tc_codel_xstats st = { ++ .maxpacket = q->stats.maxpacket, ++ .count = q->vars.count, ++ .lastcount = q->vars.lastcount, ++ .drop_overlimit = q->drop_overlimit, ++ .ldelay = codel_time_to_us(q->vars.ldelay), ++ .dropping = q->vars.dropping, ++ .ecn_mark = q->stats.ecn_mark, ++ }; ++ ++ if (q->vars.dropping) { ++ codel_tdiff_t delta = q->vars.drop_next - codel_get_time(); ++ ++ if (delta >= 0) ++ st.drop_next = codel_time_to_us(delta); ++ else ++ st.drop_next = -codel_time_to_us(-delta); ++ } ++ ++ return gnet_stats_copy_app(d, &st, sizeof(st)); ++} ++ ++static void codel_reset(struct Qdisc *sch) ++{ ++ struct codel_sched_data *q = qdisc_priv(sch); ++ ++ qdisc_reset_queue(sch); ++ codel_vars_init(&q->vars); ++} ++ ++static struct Qdisc_ops codel_qdisc_ops __read_mostly = { ++ .id = "codel", ++ .priv_size = sizeof(struct codel_sched_data), ++ ++ .enqueue = codel_qdisc_enqueue, ++ .dequeue = codel_qdisc_dequeue, ++ .peek = qdisc_peek_dequeued, ++ .init = codel_init, ++ .reset = codel_reset, ++ .change = codel_change, ++ .dump = codel_dump, ++ .dump_stats = codel_dump_stats, ++ .owner = THIS_MODULE, ++}; ++ ++static int __init codel_module_init(void) ++{ ++ return register_qdisc(&codel_qdisc_ops); ++} ++ ++static void __exit codel_module_exit(void) ++{ ++ unregister_qdisc(&codel_qdisc_ops); ++} ++ ++module_init(codel_module_init) ++module_exit(codel_module_exit) ++ ++MODULE_DESCRIPTION("Controlled Delay queue discipline"); ++MODULE_AUTHOR("Dave Taht"); ++MODULE_AUTHOR("Eric Dumazet"); ++MODULE_LICENSE("Dual BSD/GPL"); +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch b/src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch new file mode 100644 index 0000000..954d050 --- /dev/null +++ b/src/patches/linux-3.2-codel/0002-codel-use-Newton-method-instead-of-sqrt-and-divides.patch @@ -0,0 +1,189 @@ +From: Eric Dumazet <edumazet(a)google.com> +Date: Sat, 12 May 2012 03:32:13 +0000 +Subject: [PATCH 2/7] codel: use Newton method instead of sqrt() and divides + +commit 536edd67109df5e0cdb2c4ee759e9bade7976367 upstream. + +As Van pointed out, interval/sqrt(count) can be implemented using +multiplies only. + +http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_methods_for_reciprocal_square_roots + +This patch implements the Newton method and reciprocal divide. + +Total cost is 15 cycles instead of 120 on my Corei5 machine (64bit +kernel). + +There is a small 'error' for count values < 5, but we don't really care. + +I reuse a hole in struct codel_vars : + - pack the dropping boolean into one bit + - use 31bit to store the reciprocal value of sqrt(count). + +Suggested-by: Van Jacobson <van(a)pollere.net> +Signed-off-by: Eric Dumazet <edumazet(a)google.com> +Cc: Dave Taht <dave.taht(a)bufferbloat.net> +Cc: Kathleen Nichols <nichols(a)pollere.com> +Cc: Tom Herbert <therbert(a)google.com> +Cc: Matt Mathis <mattmathis(a)google.com> +Cc: Yuchung Cheng <ycheng(a)google.com> +Cc: Nandita Dukkipati <nanditad(a)google.com> +Cc: Stephen Hemminger <shemminger(a)vyatta.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> +--- + include/net/codel.h | 68 ++++++++++++++++++++++++++++----------------------- + 1 file changed, 37 insertions(+), 31 deletions(-) + +diff --git a/include/net/codel.h b/include/net/codel.h +index bce2cef..bd8747c 100644 +--- a/include/net/codel.h ++++ b/include/net/codel.h +@@ -46,6 +46,7 @@ + #include <linux/skbuff.h> + #include <net/pkt_sched.h> + #include <net/inet_ecn.h> ++#include <linux/reciprocal_div.h> + + /* Controlling Queue Delay (CoDel) algorithm + * ========================================= +@@ -123,6 +124,7 @@ struct codel_params { + * entered dropping state + * @lastcount: count at entry to dropping state + * @dropping: set to true if in dropping state ++ * @rec_inv_sqrt: reciprocal value of sqrt(count) >> 1 + * @first_above_time: when we went (or will go) continuously above target + * for interval + * @drop_next: time to drop next packet, or when we dropped last +@@ -131,7 +133,8 @@ struct codel_params { + struct codel_vars { + u32 count; + u32 lastcount; +- bool dropping; ++ bool dropping:1; ++ u32 rec_inv_sqrt:31; + codel_time_t first_above_time; + codel_time_t drop_next; + codel_time_t ldelay; +@@ -158,11 +161,7 @@ static void codel_params_init(struct codel_params *params) + + static void codel_vars_init(struct codel_vars *vars) + { +- vars->drop_next = 0; +- vars->first_above_time = 0; +- vars->dropping = false; /* exit dropping state */ +- vars->count = 0; +- vars->lastcount = 0; ++ memset(vars, 0, sizeof(*vars)); + } + + static void codel_stats_init(struct codel_stats *stats) +@@ -170,38 +169,37 @@ static void codel_stats_init(struct codel_stats *stats) + stats->maxpacket = 256; + } + +-/* return interval/sqrt(x) with good precision +- * relies on int_sqrt(unsigned long x) kernel implementation ++/* ++ * http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_me… ++ * new_invsqrt = (invsqrt / 2) * (3 - count * invsqrt^2) ++ * ++ * Here, invsqrt is a fixed point number (< 1.0), 31bit mantissa) + */ +-static u32 codel_inv_sqrt(u32 _interval, u32 _x) ++static void codel_Newton_step(struct codel_vars *vars) + { +- u64 interval = _interval; +- unsigned long x = _x; ++ u32 invsqrt = vars->rec_inv_sqrt; ++ u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 31; ++ u64 val = (3LL << 31) - ((u64)vars->count * invsqrt2); + +- /* Scale operands for max precision */ +- +-#if BITS_PER_LONG == 64 +- x <<= 32; /* On 64bit arches, we can prescale x by 32bits */ +- interval <<= 16; +-#endif ++ val = (val * invsqrt) >> 32; + +- while (x < (1UL << (BITS_PER_LONG - 2))) { +- x <<= 2; +- interval <<= 1; +- } +- do_div(interval, int_sqrt(x)); +- return (u32)interval; ++ vars->rec_inv_sqrt = val; + } + ++/* ++ * CoDel control_law is t + interval/sqrt(count) ++ * We maintain in rec_inv_sqrt the reciprocal value of sqrt(count) to avoid ++ * both sqrt() and divide operation. ++ */ + static codel_time_t codel_control_law(codel_time_t t, + codel_time_t interval, +- u32 count) ++ u32 rec_inv_sqrt) + { +- return t + codel_inv_sqrt(interval, count); ++ return t + reciprocal_divide(interval, rec_inv_sqrt << 1); + } + + +-static bool codel_should_drop(struct sk_buff *skb, ++static bool codel_should_drop(const struct sk_buff *skb, + unsigned int *backlog, + struct codel_vars *vars, + struct codel_params *params, +@@ -274,14 +272,16 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + */ + while (vars->dropping && + codel_time_after_eq(now, vars->drop_next)) { +- if (++vars->count == 0) /* avoid zero divides */ +- vars->count = ~0U; ++ vars->count++; /* dont care of possible wrap ++ * since there is no more divide ++ */ ++ codel_Newton_step(vars); + if (params->ecn && INET_ECN_set_ce(skb)) { + stats->ecn_mark++; + vars->drop_next = + codel_control_law(vars->drop_next, + params->interval, +- vars->count); ++ vars->rec_inv_sqrt); + goto end; + } + qdisc_drop(skb, sch); +@@ -296,7 +296,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + vars->drop_next = + codel_control_law(vars->drop_next, + params->interval, +- vars->count); ++ vars->rec_inv_sqrt); + } + } + } +@@ -319,12 +319,18 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + if (codel_time_before(now - vars->drop_next, + 16 * params->interval)) { + vars->count = (vars->count - vars->lastcount) | 1; ++ /* we dont care if rec_inv_sqrt approximation ++ * is not very precise : ++ * Next Newton steps will correct it quadratically. ++ */ ++ codel_Newton_step(vars); + } else { + vars->count = 1; ++ vars->rec_inv_sqrt = 0x7fffffff; + } + vars->lastcount = vars->count; + vars->drop_next = codel_control_law(now, params->interval, +- vars->count); ++ vars->rec_inv_sqrt); + } + end: + return skb; +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch b/src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch new file mode 100644 index 0000000..e1d2260 --- /dev/null +++ b/src/patches/linux-3.2-codel/0003-fq_codel-Fair-Queue-Codel-AQM.patch @@ -0,0 +1,850 @@ +From: Eric Dumazet <edumazet(a)google.com> +Date: Fri, 11 May 2012 09:30:50 +0000 +Subject: [PATCH 3/7] fq_codel: Fair Queue Codel AQM + +commit 4b549a2ef4bef9965d97cbd992ba67930cd3e0fe upstream. + +Fair Queue Codel packet scheduler + +Principles : + +- Packets are classified (internal classifier or external) on flows. +- This is a Stochastic model (as we use a hash, several flows might + be hashed on same slot) +- Each flow has a CoDel managed queue. +- Flows are linked onto two (Round Robin) lists, + so that new flows have priority on old ones. + +- For a given flow, packets are not reordered (CoDel uses a FIFO) +- head drops only. +- ECN capability is on by default. +- Very low memory footprint (64 bytes per flow) + +tc qdisc ... fq_codel [ limit PACKETS ] [ flows number ] + [ target TIME ] [ interval TIME ] [ noecn ] + [ quantum BYTES ] + +defaults : 1024 flows, 10240 packets limit, quantum : device MTU + target : 5ms (CoDel default) + interval : 100ms (CoDel default) + +Impressive results on load : + +class htb 1:1 root leaf 10: prio 0 quantum 1514 rate 200000Kbit ceil 200000Kbit burst 1475b/8 mpu 0b overhead 0b cburst 1475b/8 mpu 0b overhead 0b level 0 + Sent 43304920109 bytes 33063109 pkt (dropped 0, overlimits 0 requeues 0) + rate 201691Kbit 28595pps backlog 0b 312p requeues 0 + lended: 33063109 borrowed: 0 giants: 0 + tokens: -912 ctokens: -912 + +class fq_codel 10:1735 parent 10: + (dropped 1292, overlimits 0 requeues 0) + backlog 15140b 10p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 7.1ms +class fq_codel 10:4524 parent 10: + (dropped 1291, overlimits 0 requeues 0) + backlog 16654b 11p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 7.1ms +class fq_codel 10:4e74 parent 10: + (dropped 1290, overlimits 0 requeues 0) + backlog 6056b 4p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 6.4ms dropping drop_next 92.0ms +class fq_codel 10:628a parent 10: + (dropped 1289, overlimits 0 requeues 0) + backlog 7570b 5p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 5.4ms dropping drop_next 90.9ms +class fq_codel 10:a4b3 parent 10: + (dropped 302, overlimits 0 requeues 0) + backlog 16654b 11p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 7.1ms +class fq_codel 10:c3c2 parent 10: + (dropped 1284, overlimits 0 requeues 0) + backlog 13626b 9p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 5.9ms +class fq_codel 10:d331 parent 10: + (dropped 299, overlimits 0 requeues 0) + backlog 15140b 10p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 7.0ms +class fq_codel 10:d526 parent 10: + (dropped 12160, overlimits 0 requeues 0) + backlog 35870b 211p requeues 0 + deficit 1508 count 12160 lastcount 1 ldelay 15.3ms dropping drop_next 247us +class fq_codel 10:e2c6 parent 10: + (dropped 1288, overlimits 0 requeues 0) + backlog 15140b 10p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 7.1ms +class fq_codel 10:eab5 parent 10: + (dropped 1285, overlimits 0 requeues 0) + backlog 16654b 11p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 5.9ms +class fq_codel 10:f220 parent 10: + (dropped 1289, overlimits 0 requeues 0) + backlog 15140b 10p requeues 0 + deficit 1514 count 1 lastcount 1 ldelay 7.1ms + +qdisc htb 1: root refcnt 6 r2q 10 default 1 direct_packets_stat 0 ver 3.17 + Sent 43331086547 bytes 33092812 pkt (dropped 0, overlimits 66063544 requeues 71) + rate 201697Kbit 28602pps backlog 0b 260p requeues 71 +qdisc fq_codel 10: parent 1:1 limit 10240p flows 65536 target 5.0ms interval 100.0ms ecn + Sent 43331086547 bytes 33092812 pkt (dropped 949359, overlimits 0 requeues 0) + rate 201697Kbit 28602pps backlog 189352b 260p requeues 0 + maxpacket 1514 drop_overlimit 0 new_flow_count 5582 ecn_mark 125593 + new_flows_len 0 old_flows_len 11 + +PING 172.30.42.18 (172.30.42.18) 56(84) bytes of data. +64 bytes from 172.30.42.18: icmp_req=1 ttl=64 time=0.227 ms +64 bytes from 172.30.42.18: icmp_req=2 ttl=64 time=0.165 ms +64 bytes from 172.30.42.18: icmp_req=3 ttl=64 time=0.166 ms +64 bytes from 172.30.42.18: icmp_req=4 ttl=64 time=0.151 ms +64 bytes from 172.30.42.18: icmp_req=5 ttl=64 time=0.164 ms +64 bytes from 172.30.42.18: icmp_req=6 ttl=64 time=0.172 ms +64 bytes from 172.30.42.18: icmp_req=7 ttl=64 time=0.175 ms +64 bytes from 172.30.42.18: icmp_req=8 ttl=64 time=0.183 ms +64 bytes from 172.30.42.18: icmp_req=9 ttl=64 time=0.158 ms +64 bytes from 172.30.42.18: icmp_req=10 ttl=64 time=0.200 ms + +10 packets transmitted, 10 received, 0% packet loss, time 8999ms +rtt min/avg/max/mdev = 0.151/0.176/0.227/0.022 ms + +Much better than SFQ because of priority given to new flows, and fast +path dirtying less cache lines. + +Signed-off-by: Eric Dumazet <edumazet(a)google.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> +--- + include/linux/pkt_sched.h | 54 ++++ + net/sched/Kconfig | 11 + + net/sched/Makefile | 1 + + net/sched/sch_fq_codel.c | 624 +++++++++++++++++++++++++++++++++++++++++++++ + 4 files changed, 690 insertions(+) + create mode 100644 net/sched/sch_fq_codel.c + +diff --git a/include/linux/pkt_sched.h b/include/linux/pkt_sched.h +index cde56c2..32aef0a 100644 +--- a/include/linux/pkt_sched.h ++++ b/include/linux/pkt_sched.h +@@ -681,4 +681,58 @@ struct tc_codel_xstats { + __u32 dropping; /* are we in dropping state ? */ + }; + ++/* FQ_CODEL */ ++ ++enum { ++ TCA_FQ_CODEL_UNSPEC, ++ TCA_FQ_CODEL_TARGET, ++ TCA_FQ_CODEL_LIMIT, ++ TCA_FQ_CODEL_INTERVAL, ++ TCA_FQ_CODEL_ECN, ++ TCA_FQ_CODEL_FLOWS, ++ TCA_FQ_CODEL_QUANTUM, ++ __TCA_FQ_CODEL_MAX ++}; ++ ++#define TCA_FQ_CODEL_MAX (__TCA_FQ_CODEL_MAX - 1) ++ ++enum { ++ TCA_FQ_CODEL_XSTATS_QDISC, ++ TCA_FQ_CODEL_XSTATS_CLASS, ++}; ++ ++struct tc_fq_codel_qd_stats { ++ __u32 maxpacket; /* largest packet we've seen so far */ ++ __u32 drop_overlimit; /* number of time max qdisc ++ * packet limit was hit ++ */ ++ __u32 ecn_mark; /* number of packets we ECN marked ++ * instead of being dropped ++ */ ++ __u32 new_flow_count; /* number of time packets ++ * created a 'new flow' ++ */ ++ __u32 new_flows_len; /* count of flows in new list */ ++ __u32 old_flows_len; /* count of flows in old list */ ++}; ++ ++struct tc_fq_codel_cl_stats { ++ __s32 deficit; ++ __u32 ldelay; /* in-queue delay seen by most recently ++ * dequeued packet ++ */ ++ __u32 count; ++ __u32 lastcount; ++ __u32 dropping; ++ __s32 drop_next; ++}; ++ ++struct tc_fq_codel_xstats { ++ __u32 type; ++ union { ++ struct tc_fq_codel_qd_stats qdisc_stats; ++ struct tc_fq_codel_cl_stats class_stats; ++ }; ++}; ++ + #endif +diff --git a/net/sched/Kconfig b/net/sched/Kconfig +index fadd252..e7a8976 100644 +--- a/net/sched/Kconfig ++++ b/net/sched/Kconfig +@@ -261,6 +261,17 @@ config NET_SCH_CODEL + + If unsure, say N. + ++config NET_SCH_FQ_CODEL ++ tristate "Fair Queue Controlled Delay AQM (FQ_CODEL)" ++ help ++ Say Y here if you want to use the FQ Controlled Delay (FQ_CODEL) ++ packet scheduling algorithm. ++ ++ To compile this driver as a module, choose M here: the module ++ will be called sch_fq_codel. ++ ++ If unsure, say N. ++ + config NET_SCH_INGRESS + tristate "Ingress Qdisc" + depends on NET_CLS_ACT +diff --git a/net/sched/Makefile b/net/sched/Makefile +index 30fab03..5940a19 100644 +--- a/net/sched/Makefile ++++ b/net/sched/Makefile +@@ -38,6 +38,7 @@ obj-$(CONFIG_NET_SCH_MQPRIO) += sch_mqprio.o + obj-$(CONFIG_NET_SCH_CHOKE) += sch_choke.o + obj-$(CONFIG_NET_SCH_QFQ) += sch_qfq.o + obj-$(CONFIG_NET_SCH_CODEL) += sch_codel.o ++obj-$(CONFIG_NET_SCH_FQ_CODEL) += sch_fq_codel.o + + obj-$(CONFIG_NET_CLS_U32) += cls_u32.o + obj-$(CONFIG_NET_CLS_ROUTE4) += cls_route.o +diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c +new file mode 100644 +index 0000000..a7b3754 +--- /dev/null ++++ b/net/sched/sch_fq_codel.c +@@ -0,0 +1,624 @@ ++/* ++ * Fair Queue CoDel 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. ++ * ++ * Copyright (C) 2012 Eric Dumazet <edumazet(a)google.com> ++ */ ++ ++#include <linux/module.h> ++#include <linux/types.h> ++#include <linux/kernel.h> ++#include <linux/jiffies.h> ++#include <linux/string.h> ++#include <linux/in.h> ++#include <linux/errno.h> ++#include <linux/init.h> ++#include <linux/skbuff.h> ++#include <linux/jhash.h> ++#include <linux/slab.h> ++#include <linux/vmalloc.h> ++#include <net/netlink.h> ++#include <net/pkt_sched.h> ++#include <net/flow_keys.h> ++#include <net/codel.h> ++ ++/* Fair Queue CoDel. ++ * ++ * Principles : ++ * Packets are classified (internal classifier or external) on flows. ++ * This is a Stochastic model (as we use a hash, several flows ++ * might be hashed on same slot) ++ * Each flow has a CoDel managed queue. ++ * Flows are linked onto two (Round Robin) lists, ++ * so that new flows have priority on old ones. ++ * ++ * For a given flow, packets are not reordered (CoDel uses a FIFO) ++ * head drops only. ++ * ECN capability is on by default. ++ * Low memory footprint (64 bytes per flow) ++ */ ++ ++struct fq_codel_flow { ++ struct sk_buff *head; ++ struct sk_buff *tail; ++ struct list_head flowchain; ++ int deficit; ++ u32 dropped; /* number of drops (or ECN marks) on this flow */ ++ struct codel_vars cvars; ++}; /* please try to keep this structure <= 64 bytes */ ++ ++struct fq_codel_sched_data { ++ struct tcf_proto *filter_list; /* optional external classifier */ ++ struct fq_codel_flow *flows; /* Flows table [flows_cnt] */ ++ u32 *backlogs; /* backlog table [flows_cnt] */ ++ u32 flows_cnt; /* number of flows */ ++ u32 perturbation; /* hash perturbation */ ++ u32 quantum; /* psched_mtu(qdisc_dev(sch)); */ ++ struct codel_params cparams; ++ struct codel_stats cstats; ++ u32 drop_overlimit; ++ u32 new_flow_count; ++ ++ struct list_head new_flows; /* list of new flows */ ++ struct list_head old_flows; /* list of old flows */ ++}; ++ ++static unsigned int fq_codel_hash(const struct fq_codel_sched_data *q, ++ const struct sk_buff *skb) ++{ ++ struct flow_keys keys; ++ unsigned int hash; ++ ++ skb_flow_dissect(skb, &keys); ++ hash = jhash_3words((__force u32)keys.dst, ++ (__force u32)keys.src ^ keys.ip_proto, ++ (__force u32)keys.ports, q->perturbation); ++ return ((u64)hash * q->flows_cnt) >> 32; ++} ++ ++static unsigned int fq_codel_classify(struct sk_buff *skb, struct Qdisc *sch, ++ int *qerr) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ struct tcf_result res; ++ int result; ++ ++ if (TC_H_MAJ(skb->priority) == sch->handle && ++ TC_H_MIN(skb->priority) > 0 && ++ TC_H_MIN(skb->priority) <= q->flows_cnt) ++ return TC_H_MIN(skb->priority); ++ ++ if (!q->filter_list) ++ return fq_codel_hash(q, skb) + 1; ++ ++ *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; ++ result = tc_classify(skb, q->filter_list, &res); ++ if (result >= 0) { ++#ifdef CONFIG_NET_CLS_ACT ++ switch (result) { ++ case TC_ACT_STOLEN: ++ case TC_ACT_QUEUED: ++ *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; ++ case TC_ACT_SHOT: ++ return 0; ++ } ++#endif ++ if (TC_H_MIN(res.classid) <= q->flows_cnt) ++ return TC_H_MIN(res.classid); ++ } ++ return 0; ++} ++ ++/* helper functions : might be changed when/if skb use a standard list_head */ ++ ++/* remove one skb from head of slot queue */ ++static inline struct sk_buff *dequeue_head(struct fq_codel_flow *flow) ++{ ++ struct sk_buff *skb = flow->head; ++ ++ flow->head = skb->next; ++ skb->next = NULL; ++ return skb; ++} ++ ++/* add skb to flow queue (tail add) */ ++static inline void flow_queue_add(struct fq_codel_flow *flow, ++ struct sk_buff *skb) ++{ ++ if (flow->head == NULL) ++ flow->head = skb; ++ else ++ flow->tail->next = skb; ++ flow->tail = skb; ++ skb->next = NULL; ++} ++ ++static unsigned int fq_codel_drop(struct Qdisc *sch) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ struct sk_buff *skb; ++ unsigned int maxbacklog = 0, idx = 0, i, len; ++ struct fq_codel_flow *flow; ++ ++ /* Queue is full! Find the fat flow and drop packet from it. ++ * This might sound expensive, but with 1024 flows, we scan ++ * 4KB of memory, and we dont need to handle a complex tree ++ * in fast path (packet queue/enqueue) with many cache misses. ++ */ ++ for (i = 0; i < q->flows_cnt; i++) { ++ if (q->backlogs[i] > maxbacklog) { ++ maxbacklog = q->backlogs[i]; ++ idx = i; ++ } ++ } ++ flow = &q->flows[idx]; ++ skb = dequeue_head(flow); ++ len = qdisc_pkt_len(skb); ++ q->backlogs[idx] -= len; ++ kfree_skb(skb); ++ sch->q.qlen--; ++ sch->qstats.drops++; ++ sch->qstats.backlog -= len; ++ flow->dropped++; ++ return idx; ++} ++ ++static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ unsigned int idx; ++ struct fq_codel_flow *flow; ++ int uninitialized_var(ret); ++ ++ idx = fq_codel_classify(skb, sch, &ret); ++ if (idx == 0) { ++ if (ret & __NET_XMIT_BYPASS) ++ sch->qstats.drops++; ++ kfree_skb(skb); ++ return ret; ++ } ++ idx--; ++ ++ codel_set_enqueue_time(skb); ++ flow = &q->flows[idx]; ++ flow_queue_add(flow, skb); ++ q->backlogs[idx] += qdisc_pkt_len(skb); ++ sch->qstats.backlog += qdisc_pkt_len(skb); ++ ++ if (list_empty(&flow->flowchain)) { ++ list_add_tail(&flow->flowchain, &q->new_flows); ++ codel_vars_init(&flow->cvars); ++ q->new_flow_count++; ++ flow->deficit = q->quantum; ++ flow->dropped = 0; ++ } ++ if (++sch->q.qlen < sch->limit) ++ return NET_XMIT_SUCCESS; ++ ++ q->drop_overlimit++; ++ /* Return Congestion Notification only if we dropped a packet ++ * from this flow. ++ */ ++ if (fq_codel_drop(sch) == idx) ++ return NET_XMIT_CN; ++ ++ /* As we dropped a packet, better let upper stack know this */ ++ qdisc_tree_decrease_qlen(sch, 1); ++ return NET_XMIT_SUCCESS; ++} ++ ++/* This is the specific function called from codel_dequeue() ++ * to dequeue a packet from queue. Note: backlog is handled in ++ * codel, we dont need to reduce it here. ++ */ ++static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch) ++{ ++ struct fq_codel_flow *flow; ++ struct sk_buff *skb = NULL; ++ ++ flow = container_of(vars, struct fq_codel_flow, cvars); ++ if (flow->head) { ++ skb = dequeue_head(flow); ++ sch->qstats.backlog -= qdisc_pkt_len(skb); ++ sch->q.qlen--; ++ } ++ return skb; ++} ++ ++static struct sk_buff *fq_codel_dequeue(struct Qdisc *sch) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ struct sk_buff *skb; ++ struct fq_codel_flow *flow; ++ struct list_head *head; ++ u32 prev_drop_count, prev_ecn_mark; ++ ++begin: ++ head = &q->new_flows; ++ if (list_empty(head)) { ++ head = &q->old_flows; ++ if (list_empty(head)) ++ return NULL; ++ } ++ flow = list_first_entry(head, struct fq_codel_flow, flowchain); ++ ++ if (flow->deficit <= 0) { ++ flow->deficit += q->quantum; ++ list_move_tail(&flow->flowchain, &q->old_flows); ++ goto begin; ++ } ++ ++ prev_drop_count = q->cstats.drop_count; ++ prev_ecn_mark = q->cstats.ecn_mark; ++ ++ skb = codel_dequeue(sch, &q->cparams, &flow->cvars, &q->cstats, ++ dequeue, &q->backlogs[flow - q->flows]); ++ ++ flow->dropped += q->cstats.drop_count - prev_drop_count; ++ flow->dropped += q->cstats.ecn_mark - prev_ecn_mark; ++ ++ if (!skb) { ++ /* force a pass through old_flows to prevent starvation */ ++ if ((head == &q->new_flows) && !list_empty(&q->old_flows)) ++ list_move_tail(&flow->flowchain, &q->old_flows); ++ else ++ list_del_init(&flow->flowchain); ++ goto begin; ++ } ++ qdisc_bstats_update(sch, skb); ++ flow->deficit -= qdisc_pkt_len(skb); ++ /* We cant call qdisc_tree_decrease_qlen() if our qlen is 0, ++ * or HTB crashes. Defer it for next round. ++ */ ++ if (q->cstats.drop_count && sch->q.qlen) { ++ qdisc_tree_decrease_qlen(sch, q->cstats.drop_count); ++ q->cstats.drop_count = 0; ++ } ++ return skb; ++} ++ ++static void fq_codel_reset(struct Qdisc *sch) ++{ ++ struct sk_buff *skb; ++ ++ while ((skb = fq_codel_dequeue(sch)) != NULL) ++ kfree_skb(skb); ++} ++ ++static const struct nla_policy fq_codel_policy[TCA_FQ_CODEL_MAX + 1] = { ++ [TCA_FQ_CODEL_TARGET] = { .type = NLA_U32 }, ++ [TCA_FQ_CODEL_LIMIT] = { .type = NLA_U32 }, ++ [TCA_FQ_CODEL_INTERVAL] = { .type = NLA_U32 }, ++ [TCA_FQ_CODEL_ECN] = { .type = NLA_U32 }, ++ [TCA_FQ_CODEL_FLOWS] = { .type = NLA_U32 }, ++ [TCA_FQ_CODEL_QUANTUM] = { .type = NLA_U32 }, ++}; ++ ++static int fq_codel_change(struct Qdisc *sch, struct nlattr *opt) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ struct nlattr *tb[TCA_FQ_CODEL_MAX + 1]; ++ int err; ++ ++ if (!opt) ++ return -EINVAL; ++ ++ err = nla_parse_nested(tb, TCA_FQ_CODEL_MAX, opt, fq_codel_policy); ++ if (err < 0) ++ return err; ++ if (tb[TCA_FQ_CODEL_FLOWS]) { ++ if (q->flows) ++ return -EINVAL; ++ q->flows_cnt = nla_get_u32(tb[TCA_FQ_CODEL_FLOWS]); ++ if (!q->flows_cnt || ++ q->flows_cnt > 65536) ++ return -EINVAL; ++ } ++ sch_tree_lock(sch); ++ ++ if (tb[TCA_FQ_CODEL_TARGET]) { ++ u64 target = nla_get_u32(tb[TCA_FQ_CODEL_TARGET]); ++ ++ q->cparams.target = (target * NSEC_PER_USEC) >> CODEL_SHIFT; ++ } ++ ++ if (tb[TCA_FQ_CODEL_INTERVAL]) { ++ u64 interval = nla_get_u32(tb[TCA_FQ_CODEL_INTERVAL]); ++ ++ q->cparams.interval = (interval * NSEC_PER_USEC) >> CODEL_SHIFT; ++ } ++ ++ if (tb[TCA_FQ_CODEL_LIMIT]) ++ sch->limit = nla_get_u32(tb[TCA_FQ_CODEL_LIMIT]); ++ ++ if (tb[TCA_FQ_CODEL_ECN]) ++ q->cparams.ecn = !!nla_get_u32(tb[TCA_FQ_CODEL_ECN]); ++ ++ if (tb[TCA_FQ_CODEL_QUANTUM]) ++ q->quantum = max(256U, nla_get_u32(tb[TCA_FQ_CODEL_QUANTUM])); ++ ++ while (sch->q.qlen > sch->limit) { ++ struct sk_buff *skb = fq_codel_dequeue(sch); ++ ++ kfree_skb(skb); ++ q->cstats.drop_count++; ++ } ++ qdisc_tree_decrease_qlen(sch, q->cstats.drop_count); ++ q->cstats.drop_count = 0; ++ ++ sch_tree_unlock(sch); ++ return 0; ++} ++ ++static void *fq_codel_zalloc(size_t sz) ++{ ++ void *ptr = kzalloc(sz, GFP_KERNEL | __GFP_NOWARN); ++ ++ if (!ptr) ++ ptr = vzalloc(sz); ++ return ptr; ++} ++ ++static void fq_codel_free(void *addr) ++{ ++ if (addr) { ++ if (is_vmalloc_addr(addr)) ++ vfree(addr); ++ else ++ kfree(addr); ++ } ++} ++ ++static void fq_codel_destroy(struct Qdisc *sch) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ ++ tcf_destroy_chain(&q->filter_list); ++ fq_codel_free(q->backlogs); ++ fq_codel_free(q->flows); ++} ++ ++static int fq_codel_init(struct Qdisc *sch, struct nlattr *opt) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ int i; ++ ++ sch->limit = 10*1024; ++ q->flows_cnt = 1024; ++ q->quantum = psched_mtu(qdisc_dev(sch)); ++ q->perturbation = net_random(); ++ INIT_LIST_HEAD(&q->new_flows); ++ INIT_LIST_HEAD(&q->old_flows); ++ codel_params_init(&q->cparams); ++ codel_stats_init(&q->cstats); ++ q->cparams.ecn = true; ++ ++ if (opt) { ++ int err = fq_codel_change(sch, opt); ++ if (err) ++ return err; ++ } ++ ++ if (!q->flows) { ++ q->flows = fq_codel_zalloc(q->flows_cnt * ++ sizeof(struct fq_codel_flow)); ++ if (!q->flows) ++ return -ENOMEM; ++ q->backlogs = fq_codel_zalloc(q->flows_cnt * sizeof(u32)); ++ if (!q->backlogs) { ++ fq_codel_free(q->flows); ++ return -ENOMEM; ++ } ++ for (i = 0; i < q->flows_cnt; i++) { ++ struct fq_codel_flow *flow = q->flows + i; ++ ++ INIT_LIST_HEAD(&flow->flowchain); ++ } ++ } ++ if (sch->limit >= 1) ++ sch->flags |= TCQ_F_CAN_BYPASS; ++ else ++ sch->flags &= ~TCQ_F_CAN_BYPASS; ++ return 0; ++} ++ ++static int fq_codel_dump(struct Qdisc *sch, struct sk_buff *skb) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ struct nlattr *opts; ++ ++ opts = nla_nest_start(skb, TCA_OPTIONS); ++ if (opts == NULL) ++ goto nla_put_failure; ++ ++ if (nla_put_u32(skb, TCA_FQ_CODEL_TARGET, ++ codel_time_to_us(q->cparams.target)) || ++ nla_put_u32(skb, TCA_FQ_CODEL_LIMIT, ++ sch->limit) || ++ nla_put_u32(skb, TCA_FQ_CODEL_INTERVAL, ++ codel_time_to_us(q->cparams.interval)) || ++ nla_put_u32(skb, TCA_FQ_CODEL_ECN, ++ q->cparams.ecn) || ++ nla_put_u32(skb, TCA_FQ_CODEL_QUANTUM, ++ q->quantum) || ++ nla_put_u32(skb, TCA_FQ_CODEL_FLOWS, ++ q->flows_cnt)) ++ goto nla_put_failure; ++ ++ nla_nest_end(skb, opts); ++ return skb->len; ++ ++nla_put_failure: ++ return -1; ++} ++ ++static int fq_codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ struct tc_fq_codel_xstats st = { ++ .type = TCA_FQ_CODEL_XSTATS_QDISC, ++ .qdisc_stats.maxpacket = q->cstats.maxpacket, ++ .qdisc_stats.drop_overlimit = q->drop_overlimit, ++ .qdisc_stats.ecn_mark = q->cstats.ecn_mark, ++ .qdisc_stats.new_flow_count = q->new_flow_count, ++ }; ++ struct list_head *pos; ++ ++ list_for_each(pos, &q->new_flows) ++ st.qdisc_stats.new_flows_len++; ++ ++ list_for_each(pos, &q->old_flows) ++ st.qdisc_stats.old_flows_len++; ++ ++ return gnet_stats_copy_app(d, &st, sizeof(st)); ++} ++ ++static struct Qdisc *fq_codel_leaf(struct Qdisc *sch, unsigned long arg) ++{ ++ return NULL; ++} ++ ++static unsigned long fq_codel_get(struct Qdisc *sch, u32 classid) ++{ ++ return 0; ++} ++ ++static unsigned long fq_codel_bind(struct Qdisc *sch, unsigned long parent, ++ u32 classid) ++{ ++ /* we cannot bypass queue discipline anymore */ ++ sch->flags &= ~TCQ_F_CAN_BYPASS; ++ return 0; ++} ++ ++static void fq_codel_put(struct Qdisc *q, unsigned long cl) ++{ ++} ++ ++static struct tcf_proto **fq_codel_find_tcf(struct Qdisc *sch, unsigned long cl) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ ++ if (cl) ++ return NULL; ++ return &q->filter_list; ++} ++ ++static int fq_codel_dump_class(struct Qdisc *sch, unsigned long cl, ++ struct sk_buff *skb, struct tcmsg *tcm) ++{ ++ tcm->tcm_handle |= TC_H_MIN(cl); ++ return 0; ++} ++ ++static int fq_codel_dump_class_stats(struct Qdisc *sch, unsigned long cl, ++ struct gnet_dump *d) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ u32 idx = cl - 1; ++ struct gnet_stats_queue qs = { 0 }; ++ struct tc_fq_codel_xstats xstats; ++ ++ if (idx < q->flows_cnt) { ++ const struct fq_codel_flow *flow = &q->flows[idx]; ++ const struct sk_buff *skb = flow->head; ++ ++ memset(&xstats, 0, sizeof(xstats)); ++ xstats.type = TCA_FQ_CODEL_XSTATS_CLASS; ++ xstats.class_stats.deficit = flow->deficit; ++ xstats.class_stats.ldelay = ++ codel_time_to_us(flow->cvars.ldelay); ++ xstats.class_stats.count = flow->cvars.count; ++ xstats.class_stats.lastcount = flow->cvars.lastcount; ++ xstats.class_stats.dropping = flow->cvars.dropping; ++ if (flow->cvars.dropping) { ++ codel_tdiff_t delta = flow->cvars.drop_next - ++ codel_get_time(); ++ ++ xstats.class_stats.drop_next = (delta >= 0) ? ++ codel_time_to_us(delta) : ++ -codel_time_to_us(-delta); ++ } ++ while (skb) { ++ qs.qlen++; ++ skb = skb->next; ++ } ++ qs.backlog = q->backlogs[idx]; ++ qs.drops = flow->dropped; ++ } ++ if (gnet_stats_copy_queue(d, &qs) < 0) ++ return -1; ++ if (idx < q->flows_cnt) ++ return gnet_stats_copy_app(d, &xstats, sizeof(xstats)); ++ return 0; ++} ++ ++static void fq_codel_walk(struct Qdisc *sch, struct qdisc_walker *arg) ++{ ++ struct fq_codel_sched_data *q = qdisc_priv(sch); ++ unsigned int i; ++ ++ if (arg->stop) ++ return; ++ ++ for (i = 0; i < q->flows_cnt; i++) { ++ if (list_empty(&q->flows[i].flowchain) || ++ arg->count < arg->skip) { ++ arg->count++; ++ continue; ++ } ++ if (arg->fn(sch, i + 1, arg) < 0) { ++ arg->stop = 1; ++ break; ++ } ++ arg->count++; ++ } ++} ++ ++static const struct Qdisc_class_ops fq_codel_class_ops = { ++ .leaf = fq_codel_leaf, ++ .get = fq_codel_get, ++ .put = fq_codel_put, ++ .tcf_chain = fq_codel_find_tcf, ++ .bind_tcf = fq_codel_bind, ++ .unbind_tcf = fq_codel_put, ++ .dump = fq_codel_dump_class, ++ .dump_stats = fq_codel_dump_class_stats, ++ .walk = fq_codel_walk, ++}; ++ ++static struct Qdisc_ops fq_codel_qdisc_ops __read_mostly = { ++ .cl_ops = &fq_codel_class_ops, ++ .id = "fq_codel", ++ .priv_size = sizeof(struct fq_codel_sched_data), ++ .enqueue = fq_codel_enqueue, ++ .dequeue = fq_codel_dequeue, ++ .peek = qdisc_peek_dequeued, ++ .drop = fq_codel_drop, ++ .init = fq_codel_init, ++ .reset = fq_codel_reset, ++ .destroy = fq_codel_destroy, ++ .change = fq_codel_change, ++ .dump = fq_codel_dump, ++ .dump_stats = fq_codel_dump_stats, ++ .owner = THIS_MODULE, ++}; ++ ++static int __init fq_codel_module_init(void) ++{ ++ return register_qdisc(&fq_codel_qdisc_ops); ++} ++ ++static void __exit fq_codel_module_exit(void) ++{ ++ unregister_qdisc(&fq_codel_qdisc_ops); ++} ++ ++module_init(fq_codel_module_init) ++module_exit(fq_codel_module_exit) ++MODULE_AUTHOR("Eric Dumazet"); ++MODULE_LICENSE("GPL"); +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch b/src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch new file mode 100644 index 0000000..0d3a886 --- /dev/null +++ b/src/patches/linux-3.2-codel/0004-net-codel-Add-missing-include-linux-prefetch.h.patch @@ -0,0 +1,37 @@ +From: Geert Uytterhoeven <geert(a)linux-m68k.org> +Date: Mon, 14 May 2012 09:47:05 +0000 +Subject: [PATCH 4/7] net/codel: Add missing #include <linux/prefetch.h> +MIME-Version: 1.0 +Content-Type: text/plain; charset=UTF-8 +Content-Transfer-Encoding: 8bit + +commit ce5b4b977127ee20c3f9c3fd3637cd3796f649f5 upstream. + +m68k allmodconfig: + +net/sched/sch_codel.c: In function ‘dequeue’: +net/sched/sch_codel.c:70: error: implicit declaration of function ‘prefetch’ +make[1]: *** [net/sched/sch_codel.o] Error 1 + +Signed-off-by: Geert Uytterhoeven <geert(a)linux-m68k.org> +Acked-by: Eric Dumazet <edumazet(a)google.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> +--- + net/sched/sch_codel.c | 1 + + 1 file changed, 1 insertion(+) + +diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c +index b4a1a81..213ef60 100644 +--- a/net/sched/sch_codel.c ++++ b/net/sched/sch_codel.c +@@ -46,6 +46,7 @@ + #include <linux/kernel.h> + #include <linux/errno.h> + #include <linux/skbuff.h> ++#include <linux/prefetch.h> + #include <net/pkt_sched.h> + #include <net/codel.h> + +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch b/src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch new file mode 100644 index 0000000..728c12c --- /dev/null +++ b/src/patches/linux-3.2-codel/0005-net-codel-fix-build-errors.patch @@ -0,0 +1,55 @@ +From: Sasha Levin <levinsasha928(a)gmail.com> +Date: Mon, 14 May 2012 11:57:06 +0000 +Subject: [PATCH 5/7] net: codel: fix build errors + +commit 669d67bf777def468970f2dcba1537edf3b2d329 upstream. + +Fix the following build error: + +net/sched/sch_fq_codel.c: In function 'fq_codel_dump_stats': +net/sched/sch_fq_codel.c:464:3: error: unknown field 'qdisc_stats' specified in initializer +net/sched/sch_fq_codel.c:464:3: warning: missing braces around initializer +net/sched/sch_fq_codel.c:464:3: warning: (near initialization for 'st.<anonymous>') +net/sched/sch_fq_codel.c:465:3: error: unknown field 'qdisc_stats' specified in initializer +net/sched/sch_fq_codel.c:465:3: warning: excess elements in struct initializer +net/sched/sch_fq_codel.c:465:3: warning: (near initialization for 'st') +net/sched/sch_fq_codel.c:466:3: error: unknown field 'qdisc_stats' specified in initializer +net/sched/sch_fq_codel.c:466:3: warning: excess elements in struct initializer +net/sched/sch_fq_codel.c:466:3: warning: (near initialization for 'st') +net/sched/sch_fq_codel.c:467:3: error: unknown field 'qdisc_stats' specified in initializer +net/sched/sch_fq_codel.c:467:3: warning: excess elements in struct initializer +net/sched/sch_fq_codel.c:467:3: warning: (near initialization for 'st') +make[1]: *** [net/sched/sch_fq_codel.o] Error 1 + +Signed-off-by: Sasha Levin <levinsasha928(a)gmail.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> +--- + net/sched/sch_fq_codel.c | 9 +++++---- + 1 file changed, 5 insertions(+), 4 deletions(-) + +diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c +index a7b3754..337ff20 100644 +--- a/net/sched/sch_fq_codel.c ++++ b/net/sched/sch_fq_codel.c +@@ -461,13 +461,14 @@ static int fq_codel_dump_stats(struct Qdisc *sch, struct gnet_dump *d) + struct fq_codel_sched_data *q = qdisc_priv(sch); + struct tc_fq_codel_xstats st = { + .type = TCA_FQ_CODEL_XSTATS_QDISC, +- .qdisc_stats.maxpacket = q->cstats.maxpacket, +- .qdisc_stats.drop_overlimit = q->drop_overlimit, +- .qdisc_stats.ecn_mark = q->cstats.ecn_mark, +- .qdisc_stats.new_flow_count = q->new_flow_count, + }; + struct list_head *pos; + ++ st.qdisc_stats.maxpacket = q->cstats.maxpacket; ++ st.qdisc_stats.drop_overlimit = q->drop_overlimit; ++ st.qdisc_stats.ecn_mark = q->cstats.ecn_mark; ++ st.qdisc_stats.new_flow_count = q->new_flow_count; ++ + list_for_each(pos, &q->new_flows) + st.qdisc_stats.new_flows_len++; + +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch b/src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch new file mode 100644 index 0000000..67bb750 --- /dev/null +++ b/src/patches/linux-3.2-codel/0006-codel-use-u16-field-instead-of-31bits-for-rec_inv_sq.patch @@ -0,0 +1,90 @@ +From: Eric Dumazet <eric.dumazet(a)gmail.com> +Date: Sat, 12 May 2012 21:23:23 +0000 +Subject: [PATCH 6/7] codel: use u16 field instead of 31bits for rec_inv_sqrt + +commit 6ff272c9ad65eda219cd975b9da2dbc31cc812ee upstream. + +David pointed out gcc might generate poor code with 31bit fields. + +Using u16 is more than enough and permits a better code output. + +Also make the code intent more readable using constants, fixed point arithmetic +not being trivial for everybody. + +Suggested-by: David Miller <davem(a)davemloft.net> +Signed-off-by: Eric Dumazet <edumazet(a)google.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> +--- + include/net/codel.h | 25 +++++++++++++++---------- + 1 file changed, 15 insertions(+), 10 deletions(-) + +diff --git a/include/net/codel.h b/include/net/codel.h +index bd8747c..7546517 100644 +--- a/include/net/codel.h ++++ b/include/net/codel.h +@@ -133,13 +133,17 @@ struct codel_params { + struct codel_vars { + u32 count; + u32 lastcount; +- bool dropping:1; +- u32 rec_inv_sqrt:31; ++ bool dropping; ++ u16 rec_inv_sqrt; + codel_time_t first_above_time; + codel_time_t drop_next; + codel_time_t ldelay; + }; + ++#define REC_INV_SQRT_BITS (8 * sizeof(u16)) /* or sizeof_in_bits(rec_inv_sqrt) */ ++/* needed shift to get a Q0.32 number from rec_inv_sqrt */ ++#define REC_INV_SQRT_SHIFT (32 - REC_INV_SQRT_BITS) ++ + /** + * struct codel_stats - contains codel shared variables and stats + * @maxpacket: largest packet we've seen so far +@@ -173,17 +177,18 @@ static void codel_stats_init(struct codel_stats *stats) + * http://en.wikipedia.org/wiki/Methods_of_computing_square_roots#Iterative_me… + * new_invsqrt = (invsqrt / 2) * (3 - count * invsqrt^2) + * +- * Here, invsqrt is a fixed point number (< 1.0), 31bit mantissa) ++ * Here, invsqrt is a fixed point number (< 1.0), 32bit mantissa, aka Q0.32 + */ + static void codel_Newton_step(struct codel_vars *vars) + { +- u32 invsqrt = vars->rec_inv_sqrt; +- u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 31; +- u64 val = (3LL << 31) - ((u64)vars->count * invsqrt2); ++ u32 invsqrt = ((u32)vars->rec_inv_sqrt) << REC_INV_SQRT_SHIFT; ++ u32 invsqrt2 = ((u64)invsqrt * invsqrt) >> 32; ++ u64 val = (3LL << 32) - ((u64)vars->count * invsqrt2); + +- val = (val * invsqrt) >> 32; ++ val >>= 2; /* avoid overflow in following multiply */ ++ val = (val * invsqrt) >> (32 - 2 + 1); + +- vars->rec_inv_sqrt = val; ++ vars->rec_inv_sqrt = val >> REC_INV_SQRT_SHIFT; + } + + /* +@@ -195,7 +200,7 @@ static codel_time_t codel_control_law(codel_time_t t, + codel_time_t interval, + u32 rec_inv_sqrt) + { +- return t + reciprocal_divide(interval, rec_inv_sqrt << 1); ++ return t + reciprocal_divide(interval, rec_inv_sqrt << REC_INV_SQRT_SHIFT); + } + + +@@ -326,7 +331,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + codel_Newton_step(vars); + } else { + vars->count = 1; +- vars->rec_inv_sqrt = 0x7fffffff; ++ vars->rec_inv_sqrt = ~0U >> REC_INV_SQRT_SHIFT; + } + vars->lastcount = vars->count; + vars->drop_next = codel_control_law(now, params->interval, +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch b/src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch new file mode 100644 index 0000000..0195fdf --- /dev/null +++ b/src/patches/linux-3.2-codel/0007-fq_codel-should-use-qdisc-backlog-as-threshold.patch @@ -0,0 +1,140 @@ +From: Eric Dumazet <edumazet(a)google.com> +Date: Wed, 16 May 2012 04:39:09 +0000 +Subject: [PATCH 7/7] fq_codel: should use qdisc backlog as threshold + +commit 865ec5523dadbedefbc5710a68969f686a28d928 upstream. + +codel_should_drop() logic allows a packet being not dropped if queue +size is under max packet size. + +In fq_codel, we have two possible backlogs : The qdisc global one, and +the flow local one. + +The meaningful one for codel_should_drop() should be the global backlog, +not the per flow one, so that thin flows can have a non zero drop/mark +probability. + +Signed-off-by: Eric Dumazet <edumazet(a)google.com> +Cc: Dave Taht <dave.taht(a)bufferbloat.net> +Cc: Kathleen Nichols <nichols(a)pollere.com> +Cc: Van Jacobson <van(a)pollere.net> +Signed-off-by: David S. Miller <davem(a)davemloft.net> +--- + include/net/codel.h | 15 +++++++-------- + net/sched/sch_codel.c | 4 ++-- + net/sched/sch_fq_codel.c | 5 +++-- + 3 files changed, 12 insertions(+), 12 deletions(-) + +diff --git a/include/net/codel.h b/include/net/codel.h +index 7546517..550debf 100644 +--- a/include/net/codel.h ++++ b/include/net/codel.h +@@ -205,7 +205,7 @@ static codel_time_t codel_control_law(codel_time_t t, + + + static bool codel_should_drop(const struct sk_buff *skb, +- unsigned int *backlog, ++ struct Qdisc *sch, + struct codel_vars *vars, + struct codel_params *params, + struct codel_stats *stats, +@@ -219,13 +219,13 @@ static bool codel_should_drop(const struct sk_buff *skb, + } + + vars->ldelay = now - codel_get_enqueue_time(skb); +- *backlog -= qdisc_pkt_len(skb); ++ sch->qstats.backlog -= qdisc_pkt_len(skb); + + if (unlikely(qdisc_pkt_len(skb) > stats->maxpacket)) + stats->maxpacket = qdisc_pkt_len(skb); + + if (codel_time_before(vars->ldelay, params->target) || +- *backlog <= stats->maxpacket) { ++ sch->qstats.backlog <= stats->maxpacket) { + /* went below - stay below for at least interval */ + vars->first_above_time = 0; + return false; +@@ -249,8 +249,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + struct codel_params *params, + struct codel_vars *vars, + struct codel_stats *stats, +- codel_skb_dequeue_t dequeue_func, +- u32 *backlog) ++ codel_skb_dequeue_t dequeue_func) + { + struct sk_buff *skb = dequeue_func(vars, sch); + codel_time_t now; +@@ -261,7 +260,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + return skb; + } + now = codel_get_time(); +- drop = codel_should_drop(skb, backlog, vars, params, stats, now); ++ drop = codel_should_drop(skb, sch, vars, params, stats, now); + if (vars->dropping) { + if (!drop) { + /* sojourn time below target - leave dropping state */ +@@ -292,7 +291,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + qdisc_drop(skb, sch); + stats->drop_count++; + skb = dequeue_func(vars, sch); +- if (!codel_should_drop(skb, backlog, ++ if (!codel_should_drop(skb, sch, + vars, params, stats, now)) { + /* leave dropping state */ + vars->dropping = false; +@@ -313,7 +312,7 @@ static struct sk_buff *codel_dequeue(struct Qdisc *sch, + stats->drop_count++; + + skb = dequeue_func(vars, sch); +- drop = codel_should_drop(skb, backlog, vars, params, ++ drop = codel_should_drop(skb, sch, vars, params, + stats, now); + } + vars->dropping = true; +diff --git a/net/sched/sch_codel.c b/net/sched/sch_codel.c +index 213ef60..2f9ab17 100644 +--- a/net/sched/sch_codel.c ++++ b/net/sched/sch_codel.c +@@ -77,8 +77,8 @@ static struct sk_buff *codel_qdisc_dequeue(struct Qdisc *sch) + struct codel_sched_data *q = qdisc_priv(sch); + struct sk_buff *skb; + +- skb = codel_dequeue(sch, &q->params, &q->vars, &q->stats, +- dequeue, &sch->qstats.backlog); ++ skb = codel_dequeue(sch, &q->params, &q->vars, &q->stats, dequeue); ++ + /* We cant call qdisc_tree_decrease_qlen() if our qlen is 0, + * or HTB crashes. Defer it for next round. + */ +diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c +index 337ff20..9fc1c62 100644 +--- a/net/sched/sch_fq_codel.c ++++ b/net/sched/sch_fq_codel.c +@@ -217,13 +217,14 @@ static int fq_codel_enqueue(struct sk_buff *skb, struct Qdisc *sch) + */ + static struct sk_buff *dequeue(struct codel_vars *vars, struct Qdisc *sch) + { ++ struct fq_codel_sched_data *q = qdisc_priv(sch); + struct fq_codel_flow *flow; + struct sk_buff *skb = NULL; + + flow = container_of(vars, struct fq_codel_flow, cvars); + if (flow->head) { + skb = dequeue_head(flow); +- sch->qstats.backlog -= qdisc_pkt_len(skb); ++ q->backlogs[flow - q->flows] -= qdisc_pkt_len(skb); + sch->q.qlen--; + } + return skb; +@@ -256,7 +257,7 @@ begin: + prev_ecn_mark = q->cstats.ecn_mark; + + skb = codel_dequeue(sch, &q->cparams, &flow->cvars, &q->cstats, +- dequeue, &q->backlogs[flow - q->flows]); ++ dequeue); + + flow->dropped += q->cstats.drop_count - prev_drop_count; + flow->dropped += q->cstats.ecn_mark - prev_ecn_mark; +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch b/src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch new file mode 100644 index 0000000..c88f253 --- /dev/null +++ b/src/patches/linux-3.2-codel/0008-flow_dissector-use-a-64bit-load-store.patch @@ -0,0 +1,90 @@ +From: Eric Dumazet <eric.dumazet(a)gmail.com> +Date: Mon, 28 Nov 2011 20:30:35 +0000 +Subject: [PATCH 2/3] flow_dissector: use a 64bit load/store +MIME-Version: 1.0 +Content-Type: text/plain; charset=UTF-8 +Content-Transfer-Encoding: 8bit + +commit 4d77d2b567ec66a443792d99e96ac760991d80d0 upstream. + +Le lundi 28 novembre 2011 à 19:06 -0500, David Miller a écrit : +> From: Dimitris Michailidis <dm(a)chelsio.com> +> Date: Mon, 28 Nov 2011 08:25:39 -0800 +> +> >> +bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys +> >> *flow) +> >> +{ +> >> + int poff, nhoff = skb_network_offset(skb); +> >> + u8 ip_proto; +> >> + u16 proto = skb->protocol; +> > +> > __be16 instead of u16 for proto? +> +> I'll take care of this when I apply these patches. + +( CC trimmed ) + +Thanks David ! + +Here is a small patch to use one 64bit load/store on x86_64 instead of +two 32bit load/stores. + +[PATCH net-next] flow_dissector: use a 64bit load/store + +gcc compiler is smart enough to use a single load/store if we +memcpy(dptr, sptr, 8) on x86_64, regardless of +CONFIG_CC_OPTIMIZE_FOR_SIZE + +In IP header, daddr immediately follows saddr, this wont change in the +future. We only need to make sure our flow_keys (src,dst) fields wont +break the rule. + +Signed-off-by: Eric Dumazet <eric.dumazet(a)gmail.com> +Signed-off-by: David S. Miller <davem(a)davemloft.net> + +diff --git a/include/net/flow_keys.h b/include/net/flow_keys.h +index e4cb285..80461c1 100644 +--- a/include/net/flow_keys.h ++++ b/include/net/flow_keys.h +@@ -2,6 +2,7 @@ + #define _NET_FLOW_KEYS_H + + struct flow_keys { ++ /* (src,dst) must be grouped, in the same way than in IP header */ + __be32 src; + __be32 dst; + union { +diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c +index f0516d9..0985b9b 100644 +--- a/net/core/flow_dissector.c ++++ b/net/core/flow_dissector.c +@@ -8,6 +8,16 @@ + #include <linux/ppp_defs.h> + #include <net/flow_keys.h> + ++/* copy saddr & daddr, possibly using 64bit load/store ++ * Equivalent to : flow->src = iph->saddr; ++ * flow->dst = iph->daddr; ++ */ ++static void iph_to_flow_copy_addrs(struct flow_keys *flow, const struct iphdr *iph) ++{ ++ BUILD_BUG_ON(offsetof(typeof(*flow), dst) != ++ offsetof(typeof(*flow), src) + sizeof(flow->src)); ++ memcpy(&flow->src, &iph->saddr, sizeof(flow->src) + sizeof(flow->dst)); ++} + + bool skb_flow_dissect(const struct sk_buff *skb, struct flow_keys *flow) + { +@@ -31,8 +41,7 @@ ip: + ip_proto = 0; + else + ip_proto = iph->protocol; +- flow->src = iph->saddr; +- flow->dst = iph->daddr; ++ iph_to_flow_copy_addrs(flow, iph); + nhoff += iph->ihl * 4; + break; + } +-- +1.7.10 + diff --git a/src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch b/src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch new file mode 100644 index 0000000..ae4521d --- /dev/null +++ b/src/patches/linux-3.2-codel/0009-net-flow_dissector.c-missing-include-linux-export.h.patch @@ -0,0 +1,25 @@ +From: Jesper Dangaard Brouer <hawk(a)comx.dk> +Date: Tue, 24 Jan 2012 16:03:33 -0500 +Subject: [PATCH 3/3] net: flow_dissector.c missing include linux/export.h + +commit c452ed70771cea3af73d21a5914989137fbd28b8 upstream. + +The file net/core/flow_dissector.c seems to be missing +including linux/export.h. + +Signed-off-by: Jesper Dangaard Brouer <hawk(a)comx.dk> +Signed-off-by: David S. Miller <davem(a)davemloft.net> + +diff --git a/net/core/flow_dissector.c b/net/core/flow_dissector.c +index 0985b9b..a225089 100644 +--- a/net/core/flow_dissector.c ++++ b/net/core/flow_dissector.c +@@ -1,4 +1,5 @@ + #include <linux/skbuff.h> ++#include <linux/export.h> + #include <linux/ip.h> + #include <linux/ipv6.h> + #include <linux/if_vlan.h> +-- +1.7.10 + hooks/post-receive -- IPFire 2.x development tree

1 0

[git.ipfire.org] IPFire 3.x development tree branch, master, updated. 5425c9c9792c144a3f8a38df170303d0a0d41ee0
by git＠ipfire.org 13 Aug '12

13 Aug '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "IPFire 3.x development tree". The branch, master has been updated via 5425c9c9792c144a3f8a38df170303d0a0d41ee0 (commit) from a33f4021f466013d38bcae35b18bf8a07b50a250 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 5425c9c9792c144a3f8a38df170303d0a0d41ee0 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Mon Aug 13 22:12:11 2012 +0200 kernel: Remove ARM versatile kernel. Discussion about that is to be found here: http://lists.ipfire.org/pipermail/development/2012-August/000051.html ----------------------------------------------------------------------- Summary of changes: kernel/kernel.nm | 26 +------------------------- 1 files changed, 1 insertions(+), 25 deletions(-) Difference in files: diff --git a/kernel/kernel.nm b/kernel/kernel.nm index 61c541c..41ccf8e 100644 --- a/kernel/kernel.nm +++ b/kernel/kernel.nm @@ -5,7 +5,7 @@ name = kernel version = 3.5.1 -release = 0.2 +release = 0.3 thisapp = linux-%{version} maintainer = Michael Tremer <michael.tremer(a)ipfire.org> @@ -90,9 +90,6 @@ build # ERROR Currently all kernel images are disabled, because they # won't build with the grsecurity patch. - # Build versatile kernel. - build_kernel_versatile = 1 - # Build a kernel for Marvell Kirkwood-based devices. build_kernel_kirkwood = 1 @@ -127,11 +124,6 @@ build kernels += legacy end - # Build ARM versatile kernel. - if "%{build_kernel_versatile}" == "1" - kernels += versatile - end - # Build ARM kirkwood kernel. if "%{build_kernel_kirkwood}" == "1" kernels += kirkwood @@ -457,22 +449,6 @@ packages end end - if "%{build_kernel_versatile}" == "1" - package kernel-versatile - template KERNEL - - kernel_name = kernel-versatile - kernel_release = %{fullver}.versatile - end - - package kernel-versatile-devel - template KERNELDEVEL - - kernel_name = kernel-versatile - kernel_release = %{fullver}.versatile - end - end - if "%{build_kernel_kirkwood}" == "1" package kernel-kirkwood template KERNEL hooks/post-receive -- IPFire 3.x development tree

1 0

[git.ipfire.org] IPFire 2.x development tree branch, kernel-update, updated. 21df33072c3592bc445760b7c72e0e81e0b74d78
by git＠ipfire.org 13 Aug '12

13 Aug '12

1 0

[git.ipfire.org] IPFire 2.x development tree branch, kernel-update, updated. 337726bf3a0a21101ade7488264f386d54bbe20f
by git＠ipfire.org 13 Aug '12

13 Aug '12

1 0

[git.ipfire.org] IPFire 2.x development tree branch, kernel-update, updated. acb59f3a997cd39fbcc7b13df3a662533cec08c0
by git＠ipfire.org 13 Aug '12

13 Aug '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "IPFire 2.x development tree". The branch, kernel-update has been updated via acb59f3a997cd39fbcc7b13df3a662533cec08c0 (commit) via 8bec7790904905339ec70cb41ec3b56ce31b3349 (commit) via 593948a8efaa9c53547b5c44e11ed1b1c69814be (commit) via 4503e6b7ad857ceb87b7d6fe02e1952d911634bb (commit) via b431bfce486adcccac747c21882a6f735583104e (commit) via 4ad88740a073e13a69d61848af3a808ce1251af8 (commit) via 68ad7a1ab89eeac485ca85667fb43db3c6431e34 (commit) via 12a8cc1ed90b108fbbcb2d1093e754f05e98bdc1 (commit) via 14356fb908fd6062e987d61cc4c85a618e2fc180 (commit) via 2ade4613c7fe43298ea40947de6c54b07f48dfd0 (commit) via 7916a3bef82e2bb2ff4601c3f851e19bd762f70d (commit) via e6a97a0ca27877bb6396c120a7ab6ec4187dac85 (commit) via 9f0b5c9f4dc586433c8664074fcc46cefda0f666 (commit) via 9d60c9fd3d750da3c762811b30f7c23eb51a32da (commit) via 70e8a248c7d446655965f8e12868ed0c1e3c167c (commit) via b871af81ed08222d92d98a8e7576b3f7386d5e92 (commit) via 3b24acd0f33b4f803088929e5accc716e663c46f (commit) via ae2782ba1ffa3365719070c031ad59317c451f2f (commit) via b829fa10cc91535ca9b8c7708b3168dd54d34e9c (commit) via 08b4415ef2efe8a22a39f5e836099269fa024738 (commit) via 44fd6d2bb50f728b1ea772b9c184cfcca872b568 (commit) via 3090c39efd011f4da22fb076cf9fde846619c688 (commit) via 966e5b56c48c1676c0cbdf96c5ad659af7b4df56 (commit) from 47577eb38527c96d8edc712e6470f7984b43635f (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit acb59f3a997cd39fbcc7b13df3a662533cec08c0 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Mon Aug 13 14:15:04 2012 +0200 installer: add findutils to initrd. commit 8bec7790904905339ec70cb41ec3b56ce31b3349 Merge: 593948a 47577eb Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Mon Aug 13 14:14:00 2012 +0200 Merge branch 'kernel-update' of git.ipfire.org:/pub/git/ipfire-2.x into kernel-update commit 593948a8efaa9c53547b5c44e11ed1b1c69814be Merge: 4503e6b cc7e3a9 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Wed Aug 8 13:22:53 2012 +0200 Merge branch 'kernel-update' of git.ipfire.org:/pub/git/ipfire-2.x into kernel-update commit 4503e6b7ad857ceb87b7d6fe02e1952d911634bb Merge: b431bfc ad92a4b Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Wed Aug 1 08:56:54 2012 +0200 Merge remote-tracking branch 'ms/ccache-update' into kernel-update commit b431bfce486adcccac747c21882a6f735583104e Merge: 4ad8874 6665a03 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Sat Jul 21 23:54:31 2012 +0200 Merge remote-tracking branch 'origin/master' into kernel-update commit 4ad88740a073e13a69d61848af3a808ce1251af8 Merge: 68ad7a1 141c7c9 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Sat Jul 21 23:54:01 2012 +0200 Merge branch 'kernel-update' of git.ipfire.org:/pub/git/ipfire-2.x into kernel-update commit 68ad7a1ab89eeac485ca85667fb43db3c6431e34 Merge: 12a8cc1 3d08d93 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Fri Jul 20 09:08:40 2012 +0200 Merge remote-tracking branch 'ms/thirteen' into kernel-update commit 12a8cc1ed90b108fbbcb2d1093e754f05e98bdc1 Merge: 14356fb ee71790 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Fri Jul 20 09:05:13 2012 +0200 Merge remote-tracking branch 'ms/glibc-update' into kernel-update commit 14356fb908fd6062e987d61cc4c85a618e2fc180 Merge: 08b4415 2ade461 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Fri Jul 20 09:04:42 2012 +0200 Merge remote-tracking branch 'ms/strongswan-5' into kernel-update commit 2ade4613c7fe43298ea40947de6c54b07f48dfd0 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Thu Jul 19 17:20:05 2012 +0200 Add all changed files to the updater. commit 7916a3bef82e2bb2ff4601c3f851e19bd762f70d Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Thu Jul 19 16:54:05 2012 +0200 vpnmain.cgi: Reflect recent changes: vpn-watch removed. commit e6a97a0ca27877bb6396c120a7ab6ec4187dac85 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Thu Jul 19 16:51:50 2012 +0200 Remove vpn-watch. commit 9f0b5c9f4dc586433c8664074fcc46cefda0f666 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Thu Jul 19 16:46:00 2012 +0200 ipsec: Improve connection reloading. As pluto is no longer present, there is a lot to clean up. The connection rename hack is no longer needed and the whole ipsec stack can be controlled with the "ipsec" command. commit 9d60c9fd3d750da3c762811b30f7c23eb51a32da Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Jul 15 21:04:38 2012 +0200 initscripts: Don't create tmp dirs for pluto anymore. commit 70e8a248c7d446655965f8e12868ed0c1e3c167c Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Jul 15 21:04:05 2012 +0200 strongswan: Fix running check in services.cgi. Pluto does not exist anymore. Check for charon. commit b871af81ed08222d92d98a8e7576b3f7386d5e92 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Jul 15 15:44:17 2012 +0200 Disable vpn-watch. commit 3b24acd0f33b4f803088929e5accc716e663c46f Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Jul 15 15:43:25 2012 +0200 Create an strongswan update for preview. commit ae2782ba1ffa3365719070c031ad59317c451f2f Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Jul 15 15:34:59 2012 +0200 Update VPN CGI scripts to work with strongswan 5.0.0. Pluto is not supported anymore, the following defaults have been changed: * AES 256 is enabled by default for IKE and ESP. * DH MODP group has been set to 2048. * Compression is enabled. * IKEv2 is default. Lots of code cleanup has been done as well. commit b829fa10cc91535ca9b8c7708b3168dd54d34e9c Merge: 3090c39 b0c682c Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Jul 15 13:15:11 2012 +0200 Merge branch 'next' of ssh://git.ipfire.org/pub/git/ipfire-2.x into strongswan-5 commit 08b4415ef2efe8a22a39f5e836099269fa024738 Merge: 44fd6d2 ace40c9 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Sat Jul 7 23:17:25 2012 +0200 Merge remote-tracking branch 'origin/next' into kernel-update commit 44fd6d2bb50f728b1ea772b9c184cfcca872b568 Merge: 966e5b5 986a6b7 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Sat Jul 7 23:16:47 2012 +0200 Merge branch 'kernel-update' of git.ipfire.org:/pub/git/ipfire-2.x into kernel-update commit 3090c39efd011f4da22fb076cf9fde846619c688 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Wed Jul 4 15:53:19 2012 +0200 strongswan: Update to 5.0.0. This update removes pluto which is replaced by charon. commit 966e5b56c48c1676c0cbdf96c5ad659af7b4df56 Merge: 83d893a 7d97107 Author: Arne Fitzenreiter <arne_f(a)ipfire.org> Date: Wed Jul 4 10:40:34 2012 +0200 Merge remote-tracking branch 'origin/next' into kernel-update ----------------------------------------------------------------------- Summary of changes: config/rootfiles/common/stage2 | 5 +- config/rootfiles/common/strongswan | 90 +------- config/rootfiles/core/{55 => strongswan}/exclude | 0 config/rootfiles/core/strongswan/filelists/files | 6 + .../core/{53 => strongswan}/filelists/strongswan | 0 config/rootfiles/core/{52 => strongswan}/meta | 0 config/rootfiles/core/{61 => strongswan}/update.sh | 14 +- config/rootfiles/installer/findutils | 1 + html/cgi-bin/services.cgi | 2 +- html/cgi-bin/vpnmain.cgi | 187 +++++------------ lfs/strongswan | 27 ++- src/initscripts/init.d/tmpfs | 4 - src/misc-progs/ipsecctrl.c | 221 ++++++-------------- src/scripts/vpn-watch | 83 -------- 14 files changed, 166 insertions(+), 474 deletions(-) copy config/rootfiles/core/{55 => strongswan}/exclude (100%) create mode 100644 config/rootfiles/core/strongswan/filelists/files copy config/rootfiles/core/{53 => strongswan}/filelists/strongswan (100%) copy config/rootfiles/core/{52 => strongswan}/meta (100%) copy config/rootfiles/core/{61 => strongswan}/update.sh (92%) create mode 120000 config/rootfiles/installer/findutils delete mode 100755 src/scripts/vpn-watch Difference in files: diff --git a/config/rootfiles/common/stage2 b/config/rootfiles/common/stage2 index 796e0f3..e59763f 100644 --- a/config/rootfiles/common/stage2 +++ b/config/rootfiles/common/stage2 @@ -13,8 +13,8 @@ etc/hddtemp.db etc/host.conf etc/inittab etc/inputrc -#etc/ipsec.user.conf -#etc/ipsec.user.secrets +etc/ipsec.user.conf +etc/ipsec.user.secrets etc/issue etc/ld.so.conf etc/localtime @@ -75,7 +75,6 @@ usr/local/bin/setddns.pl usr/local/bin/settime usr/local/bin/timecheck #usr/local/bin/uname -usr/local/bin/vpn-watch #usr/local/include #usr/local/lib #usr/local/sbin diff --git a/config/rootfiles/common/strongswan b/config/rootfiles/common/strongswan index 4c7d558..ac368d6 100644 --- a/config/rootfiles/common/strongswan +++ b/config/rootfiles/common/strongswan @@ -13,133 +13,62 @@ etc/strongswan.conf #usr/lib/ipsec #usr/lib/ipsec/libcharon.a #usr/lib/ipsec/libcharon.la -usr/lib/ipsec/libcharon.so +#usr/lib/ipsec/libcharon.so usr/lib/ipsec/libcharon.so.0 usr/lib/ipsec/libcharon.so.0.0.0 #usr/lib/ipsec/libhydra.a #usr/lib/ipsec/libhydra.la -usr/lib/ipsec/libhydra.so +#usr/lib/ipsec/libhydra.so usr/lib/ipsec/libhydra.so.0 usr/lib/ipsec/libhydra.so.0.0.0 #usr/lib/ipsec/libstrongswan.a #usr/lib/ipsec/libstrongswan.la -usr/lib/ipsec/libstrongswan.so +#usr/lib/ipsec/libstrongswan.so usr/lib/ipsec/libstrongswan.so.0 usr/lib/ipsec/libstrongswan.so.0.0.0 #usr/lib/ipsec/plugins -#usr/lib/ipsec/plugins/libstrongswan-aes.a -#usr/lib/ipsec/plugins/libstrongswan-aes.la usr/lib/ipsec/plugins/libstrongswan-aes.so -#usr/lib/ipsec/plugins/libstrongswan-attr.a -#usr/lib/ipsec/plugins/libstrongswan-attr.la usr/lib/ipsec/plugins/libstrongswan-attr.so -#usr/lib/ipsec/plugins/libstrongswan-cmac.a -#usr/lib/ipsec/plugins/libstrongswan-cmac.la usr/lib/ipsec/plugins/libstrongswan-cmac.so -#usr/lib/ipsec/plugins/libstrongswan-constraints.a -#usr/lib/ipsec/plugins/libstrongswan-constraints.la usr/lib/ipsec/plugins/libstrongswan-constraints.so -#usr/lib/ipsec/plugins/libstrongswan-curl.a -#usr/lib/ipsec/plugins/libstrongswan-curl.la usr/lib/ipsec/plugins/libstrongswan-curl.so -#usr/lib/ipsec/plugins/libstrongswan-des.a -#usr/lib/ipsec/plugins/libstrongswan-des.la usr/lib/ipsec/plugins/libstrongswan-des.so -#usr/lib/ipsec/plugins/libstrongswan-dnskey.a -#usr/lib/ipsec/plugins/libstrongswan-dnskey.la usr/lib/ipsec/plugins/libstrongswan-dnskey.so -#usr/lib/ipsec/plugins/libstrongswan-fips-prf.a -#usr/lib/ipsec/plugins/libstrongswan-fips-prf.la usr/lib/ipsec/plugins/libstrongswan-fips-prf.so -#usr/lib/ipsec/plugins/libstrongswan-gmp.a -#usr/lib/ipsec/plugins/libstrongswan-gmp.la usr/lib/ipsec/plugins/libstrongswan-gmp.so -#usr/lib/ipsec/plugins/libstrongswan-hmac.a -#usr/lib/ipsec/plugins/libstrongswan-hmac.la usr/lib/ipsec/plugins/libstrongswan-hmac.so -#usr/lib/ipsec/plugins/libstrongswan-kernel-netlink.a -#usr/lib/ipsec/plugins/libstrongswan-kernel-netlink.la usr/lib/ipsec/plugins/libstrongswan-kernel-netlink.so -#usr/lib/ipsec/plugins/libstrongswan-md5.a -#usr/lib/ipsec/plugins/libstrongswan-md5.la usr/lib/ipsec/plugins/libstrongswan-md5.so -#usr/lib/ipsec/plugins/libstrongswan-pem.a -#usr/lib/ipsec/plugins/libstrongswan-pem.la +usr/lib/ipsec/plugins/libstrongswan-nonce.so +usr/lib/ipsec/plugins/libstrongswan-openssl.so +usr/lib/ipsec/plugins/libstrongswan-padlock.so usr/lib/ipsec/plugins/libstrongswan-pem.so -#usr/lib/ipsec/plugins/libstrongswan-pgp.a -#usr/lib/ipsec/plugins/libstrongswan-pgp.la usr/lib/ipsec/plugins/libstrongswan-pgp.so -#usr/lib/ipsec/plugins/libstrongswan-pkcs1.a -#usr/lib/ipsec/plugins/libstrongswan-pkcs1.la usr/lib/ipsec/plugins/libstrongswan-pkcs1.so -#usr/lib/ipsec/plugins/libstrongswan-pkcs8.a -#usr/lib/ipsec/plugins/libstrongswan-pkcs8.la usr/lib/ipsec/plugins/libstrongswan-pkcs8.so -#usr/lib/ipsec/plugins/libstrongswan-pubkey.a -#usr/lib/ipsec/plugins/libstrongswan-pubkey.la usr/lib/ipsec/plugins/libstrongswan-pubkey.so -#usr/lib/ipsec/plugins/libstrongswan-random.a -#usr/lib/ipsec/plugins/libstrongswan-random.la usr/lib/ipsec/plugins/libstrongswan-random.so -#usr/lib/ipsec/plugins/libstrongswan-resolve.a -#usr/lib/ipsec/plugins/libstrongswan-resolve.la usr/lib/ipsec/plugins/libstrongswan-resolve.so -#usr/lib/ipsec/plugins/libstrongswan-revocation.a -#usr/lib/ipsec/plugins/libstrongswan-revocation.la usr/lib/ipsec/plugins/libstrongswan-revocation.so -#usr/lib/ipsec/plugins/libstrongswan-sha1.a -#usr/lib/ipsec/plugins/libstrongswan-sha1.la usr/lib/ipsec/plugins/libstrongswan-sha1.so -#usr/lib/ipsec/plugins/libstrongswan-sha2.a -#usr/lib/ipsec/plugins/libstrongswan-sha2.la usr/lib/ipsec/plugins/libstrongswan-sha2.so -#usr/lib/ipsec/plugins/libstrongswan-socket-raw.a -#usr/lib/ipsec/plugins/libstrongswan-socket-raw.la -usr/lib/ipsec/plugins/libstrongswan-socket-raw.so -#usr/lib/ipsec/plugins/libstrongswan-stroke.a -#usr/lib/ipsec/plugins/libstrongswan-stroke.la +usr/lib/ipsec/plugins/libstrongswan-socket-default.so usr/lib/ipsec/plugins/libstrongswan-stroke.so -#usr/lib/ipsec/plugins/libstrongswan-updown.a -#usr/lib/ipsec/plugins/libstrongswan-updown.la usr/lib/ipsec/plugins/libstrongswan-updown.so -#usr/lib/ipsec/plugins/libstrongswan-x509.a -#usr/lib/ipsec/plugins/libstrongswan-x509.la usr/lib/ipsec/plugins/libstrongswan-x509.so -#usr/lib/ipsec/plugins/libstrongswan-xauth.a -#usr/lib/ipsec/plugins/libstrongswan-xauth.la -usr/lib/ipsec/plugins/libstrongswan-xauth.so -#usr/lib/ipsec/plugins/libstrongswan-xcbc.a -#usr/lib/ipsec/plugins/libstrongswan-xcbc.la +usr/lib/ipsec/plugins/libstrongswan-xauth-generic.so usr/lib/ipsec/plugins/libstrongswan-xcbc.so #usr/libexec/ipsec usr/libexec/ipsec/_copyright -usr/libexec/ipsec/_pluto_adns usr/libexec/ipsec/_updown usr/libexec/ipsec/_updown_espmark usr/libexec/ipsec/charon usr/libexec/ipsec/openac usr/libexec/ipsec/pki -usr/libexec/ipsec/pluto usr/libexec/ipsec/scepclient usr/libexec/ipsec/starter usr/libexec/ipsec/stroke -usr/libexec/ipsec/whack usr/sbin/ipsec -#usr/share/man/man3/anyaddr.3 -#usr/share/man/man3/atoaddr.3 -#usr/share/man/man3/atoasr.3 -#usr/share/man/man3/atoul.3 -#usr/share/man/man3/goodmask.3 -#usr/share/man/man3/initaddr.3 -#usr/share/man/man3/initsubnet.3 -#usr/share/man/man3/portof.3 -#usr/share/man/man3/rangetosubnet.3 -#usr/share/man/man3/sameaddr.3 -#usr/share/man/man3/subnetof.3 -#usr/share/man/man3/ttoaddr.3 -#usr/share/man/man3/ttodata.3 -#usr/share/man/man3/ttosa.3 -#usr/share/man/man3/ttoul.3 #usr/share/man/man5/ipsec.conf.5 #usr/share/man/man5/ipsec.secrets.5 #usr/share/man/man5/strongswan.conf.5 @@ -147,7 +76,4 @@ usr/sbin/ipsec #usr/share/man/man8/_updown_espmark.8 #usr/share/man/man8/ipsec.8 #usr/share/man/man8/openac.8 -#usr/share/man/man8/pluto.8 #usr/share/man/man8/scepclient.8 -etc/ipsec.user.conf -etc/ipsec.user.secrets diff --git a/config/rootfiles/core/strongswan/exclude b/config/rootfiles/core/strongswan/exclude new file mode 100644 index 0000000..7360266 --- /dev/null +++ b/config/rootfiles/core/strongswan/exclude @@ -0,0 +1,12 @@ +srv/web/ipfire/html/proxy.pac +etc/udev/rules.d/30-persistent-network.rules +etc/ipsec.conf +etc/ipsec.secrets +etc/ipsec.user.conf +etc/ipsec.user.secrets +var/updatecache +etc/localtime +var/ipfire/ovpn +etc/ssh/ssh_config +etc/ssh/sshd_config +etc/ssl/openssl.cnf diff --git a/config/rootfiles/core/strongswan/filelists/files b/config/rootfiles/core/strongswan/filelists/files new file mode 100644 index 0000000..4aa5440 --- /dev/null +++ b/config/rootfiles/core/strongswan/filelists/files @@ -0,0 +1,6 @@ +etc/system-release +etc/issue +etc/rc.d/init.d/tmpfs +srv/web/ipfire/cgi-bin/services.cgi +srv/web/ipfire/cgi-bin/vpnmain.cgi +usr/local/bin/ipsecctrl diff --git a/config/rootfiles/core/strongswan/filelists/strongswan b/config/rootfiles/core/strongswan/filelists/strongswan new file mode 120000 index 0000000..90c727e --- /dev/null +++ b/config/rootfiles/core/strongswan/filelists/strongswan @@ -0,0 +1 @@ +../../../common/strongswan \ No newline at end of file diff --git a/config/rootfiles/core/strongswan/meta b/config/rootfiles/core/strongswan/meta new file mode 100644 index 0000000..d547fa8 --- /dev/null +++ b/config/rootfiles/core/strongswan/meta @@ -0,0 +1 @@ +DEPS="" diff --git a/config/rootfiles/core/strongswan/update.sh b/config/rootfiles/core/strongswan/update.sh new file mode 100644 index 0000000..7ef3f2f --- /dev/null +++ b/config/rootfiles/core/strongswan/update.sh @@ -0,0 +1,89 @@ +#!/bin/bash +############################################################################ +# # +# This file is part of the IPFire Firewall. # +# # +# IPFire 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 3 of the License, or # +# (at your option) any later version. # +# # +# IPFire is distributed in the hope that it will be useful, # +# but WITHOUT ANY WARRANTY; without even the implied warranty of # +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # +# GNU General Public License for more details. # +# # +# You should have received a copy of the GNU General Public License # +# along with IPFire; if not, write to the Free Software # +# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA # +# # +# Copyright (C) 2012 IPFire-Team <info(a)ipfire.org>. # +# # +############################################################################ +# +. /opt/pakfire/lib/functions.sh +/usr/local/bin/backupctrl exclude >/dev/null 2>&1 + +# +# Remove old core updates from pakfire cache to save space... +core=61 +for (( i=1; i<=$core; i++ )) +do + rm -f /var/cache/pakfire/core-upgrade-*-$i.ipfire +done + +# +#Stop services +ipsecctrl D + +# +#Extract files +extract_files + +# Remove old pluto binaries. +rm -f /usr/libexec/ipsec/{pluto,_pluto_adns,whack} +rm -f /usr/local/bin/vpn-watch + +# +#Start services + +# Call the CGI script to regenerate the configuration files. +/srv/web/ipfire/cgi-bin/vpnmain.cgi +ipsecctrl S + +# +#Update Language cache +perl -e "require '/var/ipfire/lang.pl'; &Lang::BuildCacheLang" + +#Rebuild module dep's +#arch=`uname -m` +#if [ ${arch::3} == "arm" ]; then +# depmod -a 2.6.32.45-ipfire-versatile >/dev/null 2>&1 +# depmod -a 2.6.32.45-ipfire-kirkwood >/dev/null 2>&1 +#else +# depmod -a 2.6.32.45-ipfire >/dev/null 2>&1 +# depmod -a 2.6.32.45-ipfire-pae >/dev/null 2>&1 +# depmod -a 2.6.32.45-ipfire-xen >/dev/null 2>&1 +#fi + + +#Rebuild initrd's because some compat-wireless modules are inside +#/sbin/dracut --force --verbose /boot/ipfirerd-2.6.32.45.img 2.6.32.45-ipfire +#if [ -e /boot/ipfirerd-2.6.32.45-pae.img ]; then +#/sbin/dracut --force --verbose /boot/ipfirerd-2.6.32.45-pae.img 2.6.32.45-ipfire-pae +#fi +#if [ -e /boot/ipfirerd-2.6.32.45-xen.img ]; then +#/sbin/dracut --force --verbose /boot/ipfirerd-2.6.32.45-xen.img 2.6.32.45-ipfire-xen +#fi + +sync + +# This update need a reboot... +#touch /var/run/need_reboot + +# +#Finish +/etc/init.d/fireinfo start +sendprofile +#Don't report the exitcode last command +exit 0 diff --git a/config/rootfiles/installer/findutils b/config/rootfiles/installer/findutils new file mode 120000 index 0000000..1114c4c --- /dev/null +++ b/config/rootfiles/installer/findutils @@ -0,0 +1 @@ +../common/findutils \ No newline at end of file diff --git a/html/cgi-bin/services.cgi b/html/cgi-bin/services.cgi index 123c325..22a9ac7 100644 --- a/html/cgi-bin/services.cgi +++ b/html/cgi-bin/services.cgi @@ -54,7 +54,7 @@ my %servicenames =( $Lang::tr{'kernel logging server'} => 'klogd', $Lang::tr{'ntp server'} => 'ntpd', $Lang::tr{'secure shell server'} => 'sshd', - $Lang::tr{'vpn'} => 'pluto', + $Lang::tr{'vpn'} => 'charon', $Lang::tr{'web proxy'} => 'squid', 'OpenVPN' => 'openvpn' ); diff --git a/html/cgi-bin/vpnmain.cgi b/html/cgi-bin/vpnmain.cgi index 0fb7c93..831ef93 100644 --- a/html/cgi-bin/vpnmain.cgi +++ b/html/cgi-bin/vpnmain.cgi @@ -73,17 +73,9 @@ $cgiparams{'ENABLED'} = 'off'; $cgiparams{'EDIT_ADVANCED'} = 'off'; $cgiparams{'ACTION'} = ''; $cgiparams{'CA_NAME'} = ''; -$cgiparams{'DBG_CRYPT'} = ''; -$cgiparams{'DBG_PARSING'} = ''; -$cgiparams{'DBG_EMITTING'} = ''; -$cgiparams{'DBG_CONTROL'} = ''; -$cgiparams{'DBG_KLIPS'} = ''; -$cgiparams{'DBG_DNS'} = ''; -$cgiparams{'DBG_NAT_T'} = ''; $cgiparams{'KEY'} = ''; $cgiparams{'TYPE'} = ''; $cgiparams{'ADVANCED'} = ''; -$cgiparams{'INTERFACE'} = ''; $cgiparams{'NAME'} = ''; $cgiparams{'LOCAL_SUBNET'} = ''; $cgiparams{'REMOTE_SUBNET'} = ''; @@ -253,50 +245,8 @@ sub writeipsecfiles { flock CONF, 2; flock SECRETS, 2; print CONF "version 2\n\n"; - print CONF "config setup\n"; - #create an ipsec Interface for each 'enabled' ones - #loop trought configuration and add physical interfaces to the list - my $interfaces = "\tinterfaces=\""; - foreach my $key (keys %lconfighash) { - next if ($lconfighash{$key}[0] ne 'on'); - $interfaces .= "%defaultroute " if ($interfaces !~ /defaultroute/ && $lconfighash{$key}[26] eq 'RED'); - $interfaces .= "$netsettings{'GREEN_DEV'} " if ($interfaces !~ /ipsec1/ && $lconfighash{$key}[26] eq 'GREEN'); - $interfaces .= "$netsettings{'BLUE_DEV'} " if ($interfaces !~ /ipsec2/ && $lconfighash{$key}[26] eq 'BLUE'); - $interfaces .= "$netsettings{'ORANGE_DEV'} " if ($interfaces !~ /ipsec3/ && $lconfighash{$key}[26] eq 'ORANGE'); - } - print CONF $interfaces . "\"\n"; - - my $plutodebug = ''; # build debug list - map ($plutodebug .= $lvpnsettings{$_} eq 'on' ? lc (substr($_,4)).' ' : '', - ('DBG_CRYPT','DBG_PARSING','DBG_EMITTING','DBG_CONTROL', - 'DBG_DNS')); - $plutodebug = 'none' if $plutodebug eq ''; # if nothing selected, use 'none'. - #print CONF "\tklipsdebug=\"none\"\n"; - print CONF "\tplutodebug=\"$plutodebug\"\n"; - # deprecated in ipsec.conf version 2 - #print CONF "\tplutoload=%search\n"; - #print CONF "\tplutostart=%search\n"; - print CONF "\tuniqueids=yes\n"; - print CONF "\tnat_traversal=yes\n"; - print CONF "\toverridemtu=$lvpnsettings{'VPN_OVERRIDE_MTU'}\n" if ($lvpnsettings{'VPN_OVERRIDE_MTU'} ne ''); - print CONF "\tvirtual_private=%v4:10.0.0.0/8,%v4:172.16.0.0/12,%v4:192.168.0.0/16"; - print CONF ",%v4:!$green_cidr"; - if (length($netsettings{'ORANGE_DEV'}) > 2) { - print CONF ",%v4:!$orange_cidr"; - } - if (length($netsettings{'BLUE_DEV'}) > 2) { - print CONF ",%v4:!$blue_cidr"; - } - foreach my $key (keys %lconfighash) { - if ($lconfighash{$key}[3] eq 'net') { - print CONF ",%v4:!$lconfighash{$key}[11]"; - } - } - print CONF "\n\n"; print CONF "conn %default\n"; - print CONF "\tkeyingtries=0\n"; - #strongswan doesn't know this - #print CONF "\tdisablearrivalcheck=no\n"; + print CONF "\tkeyingtries=%forever\n"; print CONF "\n"; # Add user includes to config file @@ -329,7 +279,6 @@ sub writeipsecfiles { print CONF "conn $lconfighash{$key}[1]\n"; print CONF "\tleft=$localside\n"; - print CONF "\tleftnexthop=%defaultroute\n" if ($lconfighash{$key}[26] eq 'RED' && $lvpnsettings{'VPN_IP'} ne '%defaultroute'); my $cidr_net=&General::ipcidr($lconfighash{$key}[8]); print CONF "\tleftsubnet=$cidr_net\n"; print CONF "\tleftfirewall=yes\n"; @@ -339,7 +288,6 @@ sub writeipsecfiles { if ($lconfighash{$key}[3] eq 'net') { my $cidr_net=&General::ipcidr($lconfighash{$key}[11]); print CONF "\trightsubnet=$cidr_net\n"; - print CONF "\trightnexthop=%defaultroute\n"; } elsif ($lconfighash{$key}[10] eq '%any' && $lconfighash{$key}[14] eq 'on') { #vhost allowed for roadwarriors? print CONF "\trightsubnet=vhost:%no,%priv\n"; } @@ -354,6 +302,9 @@ sub writeipsecfiles { print CONF "\tleftid=\"$lconfighash{$key}[7]\"\n" if ($lconfighash{$key}[7]); print CONF "\trightid=\"$lconfighash{$key}[9]\"\n" if ($lconfighash{$key}[9]); + # Is PFS enabled? + my $pfs = $lconfighash{$key}[28] eq 'on' ? 'on' : 'off'; + # Algorithms if ($lconfighash{$key}[18] && $lconfighash{$key}[19] && $lconfighash{$key}[20]) { print CONF "\tike="; @@ -379,11 +330,25 @@ sub writeipsecfiles { print CONF "\tesp="; my @encs = split('\|', $lconfighash{$key}[21]); my @ints = split('\|', $lconfighash{$key}[22]); + my @groups = split('\|', $lconfighash{$key}[20]); my $comma = 0; foreach my $i (@encs) { foreach my $j (@ints) { - if ($comma != 0) { print CONF ","; } else { $comma = 1; } - print CONF "$i-$j"; + my $modp = ""; + if ($pfs eq "on") { + foreach my $k (@groups) { + if ($comma != 0) { print CONF ","; } else { $comma = 1; } + if ($pfs eq "on") { + $modp = "-modp$k"; + } else { + $modp = ""; + } + print CONF "$i-$j$modp"; + } + } else { + if ($comma != 0) { print CONF ","; } else { $comma = 1; } + print CONF "$i-$j"; + } } } if ($lconfighash{$key}[24] eq 'on') { #only proposed algorythms? @@ -392,9 +357,6 @@ sub writeipsecfiles { print CONF "\n"; } } - if ($lconfighash{$key}[23]) { - print CONF "\tpfsgroup=$lconfighash{$key}[23]\n"; - } # IKE V1 or V2 if (! $lconfighash{$key}[29]) { @@ -414,9 +376,6 @@ sub writeipsecfiles { print CONF "\tdpdtimeout=120\n"; print CONF "\tdpdaction=$lconfighash{$key}[27]\n"; - # Disable pfs ? - print CONF "\tpfs=". ($lconfighash{$key}[28] eq 'on' ? "yes\n" : "no\n"); - # Build Authentication details: LEFTid RIGHTid : PSK psk my $psk_line; if ($lconfighash{$key}[4] eq 'psk') { @@ -450,6 +409,12 @@ sub writeipsecfiles { close(SECRETS); } +# Hook to regenerate the configuration files. +if ($ENV{"REMOTE_ADDR"} eq "") { + writeipsecfiles; + exit(0); +} + ### ### Save main settings ### @@ -466,29 +431,13 @@ if ($cgiparams{'ACTION'} eq $Lang::tr{'save'} && $cgiparams{'TYPE'} eq '' && $cg goto SAVE_ERROR; } - unless ($cgiparams{'VPN_OVERRIDE_MTU'} =~ /^(|[0-9]{1,5})$/ ) { #allow 0-99999 - $errormessage = $Lang::tr{'vpn mtu invalid'}; - goto SAVE_ERROR; - } - - unless ($cgiparams{'VPN_WATCH'} =~ /^(|off|on)$/ ) { - $errormessage = $Lang::tr{'invalid input'}; - goto SAVE_ERROR; - } - if ( $cgiparams{'RW_NET'} ne '' and !&General::validipandmask($cgiparams{'RW_NET'}) ) { $errormessage = $Lang::tr{'urlfilter invalid ip or mask error'}; goto SAVE_ERROR; } - map ($vpnsettings{$_} = $cgiparams{$_}, - ('ENABLED','DBG_CRYPT','DBG_PARSING','DBG_EMITTING','DBG_CONTROL', - 'DBG_DNS')); - $vpnsettings{'VPN_IP'} = $cgiparams{'VPN_IP'}; $vpnsettings{'VPN_DELAYED_START'} = $cgiparams{'VPN_DELAYED_START'}; - $vpnsettings{'VPN_OVERRIDE_MTU'} = $cgiparams{'VPN_OVERRIDE_MTU'}; - $vpnsettings{'VPN_WATCH'} = $cgiparams{'VPN_WATCH'}; $vpnsettings{'RW_NET'} = $cgiparams{'RW_NET'}; &General::writehash("${General::swroot}/vpn/settings", \%vpnsettings); &writeipsecfiles(); @@ -1298,7 +1247,6 @@ END $cgiparams{'REMOTE'} = $confighash{$cgiparams{'KEY'}}[10]; $cgiparams{'REMOTE_SUBNET'} = $confighash{$cgiparams{'KEY'}}[11]; $cgiparams{'REMARK'} = $confighash{$cgiparams{'KEY'}}[25]; - $cgiparams{'INTERFACE'} = $confighash{$cgiparams{'KEY'}}[26]; $cgiparams{'DPD_ACTION'} = $confighash{$cgiparams{'KEY'}}[27]; $cgiparams{'IKE_VERSION'} = $confighash{$cgiparams{'KEY'}}[29]; $cgiparams{'IKE_ENCRYPTION'} = $confighash{$cgiparams{'KEY'}}[18]; @@ -1801,7 +1749,7 @@ END $confighash{$key}[9] = $cgiparams{'REMOTE_ID'}; $confighash{$key}[10] = $cgiparams{'REMOTE'}; $confighash{$key}[25] = $cgiparams{'REMARK'}; - $confighash{$key}[26] = $cgiparams{'INTERFACE'}; + $confighash{$key}[26] = ""; # Formerly INTERFACE $confighash{$key}[27] = $cgiparams{'DPD_ACTION'}; $confighash{$key}[29] = $cgiparams{'IKE_VERSION'}; @@ -1859,28 +1807,25 @@ END $cgiparams{'DPD_ACTION'} = 'restart'; } - # Default IKE Version to V1 - if (! $cgiparams{'IKE_VERSION'}) { - $cgiparams{'IKE_VERSION'} = 'ikev1'; + # Default IKE Version to v2 + if (!$cgiparams{'IKE_VERSION'}) { + $cgiparams{'IKE_VERSION'} = 'ikev2'; } - # Default is yes for 'pfs' - $cgiparams{'PFS'} = 'on'; - # ID are empty $cgiparams{'LOCAL_ID'} = ''; $cgiparams{'REMOTE_ID'} = ''; #use default advanced value - $cgiparams{'IKE_ENCRYPTION'} = 'aes128|3des'; #[18]; + $cgiparams{'IKE_ENCRYPTION'} = 'aes256|aes128|3des'; #[18]; $cgiparams{'IKE_INTEGRITY'} = 'sha|md5'; #[19]; - $cgiparams{'IKE_GROUPTYPE'} = '1536|1024'; #[20]; + $cgiparams{'IKE_GROUPTYPE'} = '2048'; #[20]; $cgiparams{'IKE_LIFETIME'} = '1'; #[16]; - $cgiparams{'ESP_ENCRYPTION'} = 'aes128|3des'; #[21]; + $cgiparams{'ESP_ENCRYPTION'} = 'aes256|aes128|3des'; #[21]; $cgiparams{'ESP_INTEGRITY'} = 'sha1|md5'; #[22]; $cgiparams{'ESP_GROUPTYPE'} = ''; #[23]; $cgiparams{'ESP_KEYLIFE'} = '8'; #[17]; - $cgiparams{'COMPRESSION'} = 'off'; #[13]; + $cgiparams{'COMPRESSION'} = 'on'; #[13]; $cgiparams{'ONLY_PROPOSED'} = 'off'; #[24]; $cgiparams{'PFS'} = 'on'; #[28]; $cgiparams{'VHOST'} = 'on'; #[14]; @@ -1903,12 +1848,6 @@ END $checked{'AUTH'}{'auth-dn'} = ''; $checked{'AUTH'}{$cgiparams{'AUTH'}} = "checked='checked'"; - $selected{'INTERFACE'}{'RED'} = ''; - $selected{'INTERFACE'}{'ORANGE'} = ''; - $selected{'INTERFACE'}{'GREEN'} = ''; - $selected{'INTERFACE'}{'BLUE'} = ''; - $selected{'INTERFACE'}{$cgiparams{'INTERFACE'}} = "selected='selected'"; - $selected{'DPD_ACTION'}{'clear'} = ''; $selected{'DPD_ACTION'}{'hold'} = ''; $selected{'DPD_ACTION'}{'restart'} = ''; @@ -1975,22 +1914,24 @@ END $blob = "<img src='/blob.gif' alt='*' />"; }; - print "<tr><td>$Lang::tr{'host ip'}:</td>"; - print "<td><select name='INTERFACE'>"; - print "<option value='RED' $selected{'INTERFACE'}{'RED'}>RED ($vpnsettings{'VPN_IP'})</option>"; - print "<option value='GREEN' $selected{'INTERFACE'}{'GREEN'}>GREEN ($netsettings{'GREEN_ADDRESS'})</option>"; - print "<option value='BLUE' $selected{'INTERFACE'}{'BLUE'}>BLUE ($netsettings{'BLUE_ADDRESS'})</option>" if ($netsettings{'BLUE_DEV'} ne ''); - print "<option value='ORANGE' $selected{'INTERFACE'}{'ORANGE'}>ORANGE ($netsettings{'ORANGE_ADDRESS'})</option>" if ($netsettings{'ORANGE_DEV'} ne ''); - print "</select></td>"; print <<END + <tr> <td class='boldbase'>$Lang::tr{'remote host/ip'}: $blob</td> - <td><input type='text' name='REMOTE' value='$cgiparams{'REMOTE'}' size='30' /></td> - </tr><tr> - <td class='boldbase' nowrap='nowrap'>$Lang::tr{'local subnet'}</td> - <td><input type='text' name='LOCAL_SUBNET' value='$cgiparams{'LOCAL_SUBNET'}' size='30' /></td> + <td> + <input type='text' name='REMOTE' value='$cgiparams{'REMOTE'}' size='30' /> + </td> <td class='boldbase' nowrap='nowrap'>$Lang::tr{'remote subnet'}</td> - <td><input $disabled type='text' name='REMOTE_SUBNET' value='$cgiparams{'REMOTE_SUBNET'}' size='30' /></td> - </tr><tr> + <td> + <input $disabled type='text' name='REMOTE_SUBNET' value='$cgiparams{'REMOTE_SUBNET'}' size='30' /> + </td> + </tr> + <tr> + <td class='boldbase' nowrap='nowrap'>$Lang::tr{'local subnet'}</td> + <td colspan='3'> + <input type='text' name='LOCAL_SUBNET' value='$cgiparams{'LOCAL_SUBNET'}' size='30' /> + </td> + </tr> + <tr> <td class='boldbase'>$Lang::tr{'vpn local id'}:<br />($Lang::tr{'eg'} <tt>@xy.example.com</tt>)</td> <td><input type='text' name='LOCAL_ID' value='$cgiparams{'LOCAL_ID'}' /></td> <td class='boldbase'>$Lang::tr{'vpn remote id'}:</td> @@ -1999,22 +1940,18 @@ END </tr><td><br /></td><tr> <td>$Lang::tr{'vpn keyexchange'}:</td> <td><select name='IKE_VERSION'> - <option value='ikev1' $selected{'IKE_VERSION'}{'ikev1'}>IKEv1</option> <option value='ikev2' $selected{'IKE_VERSION'}{'ikev2'}>IKEv2</option> - </select></a> + <option value='ikev1' $selected{'IKE_VERSION'}{'ikev1'}>IKEv1</option> + </select> </td> <td>$Lang::tr{'dpd action'}:</td> <td><select name='DPD_ACTION'> <option value='clear' $selected{'DPD_ACTION'}{'clear'}>clear</option> <option value='hold' $selected{'DPD_ACTION'}{'hold'}>hold</option> <option value='restart' $selected{'DPD_ACTION'}{'restart'}>restart</option> - </select>  <a href='http://www.openswan.com/docs/local/README.DPD'>?</a> + </select> </td> </tr><tr> - <td class='boldbase'>$Lang::tr{'remark title'} <img src='/blob.gif' alt='*' /></td> <td colspan='3'><input type='text' name='REMARK' value='$cgiparams{'REMARK'}' size='55' maxlength='50' /></td> </tr> @@ -2447,11 +2384,7 @@ EOF $cgiparams{'VPN_IP'} ='%defaultroute' if ($cgiparams{'VPN_IP'} eq ''); $cgiparams{'VPN_DELAYED_START'} = 0 if (! defined ($cgiparams{'VPN_DELAYED_START'})); - $checked{'VPN_WATCH'} = $cgiparams{'VPN_WATCH'} eq 'on' ? "checked='checked'" : '' ; - map ($checked{$_} = $cgiparams{$_} eq 'on' ? "checked='checked'" : '', - ('ENABLED','DBG_CRYPT','DBG_PARSING','DBG_EMITTING','DBG_CONTROL', - 'DBG_DNS')); - + $checked{'ENABLED'} = $cgiparams{'ENABLED'} eq 'on' ? "checked='checked'" : ''; &Header::showhttpheaders(); &Header::openpage($Lang::tr{'vpn configuration main'}, 1, ''); @@ -2475,13 +2408,6 @@ EOF </tr> END ; - print <<END - <tr> - <td class='base' nowrap='nowrap'>$Lang::tr{'override mtu'}: <img src='/blob.gif' alt='*' /></td> - <td ><input type='text' name='VPN_OVERRIDE_MTU' value='$cgiparams{'VPN_OVERRIDE_MTU'}' /></td> - </tr> -END - ; print <<END <tr> <td class='base' nowrap='nowrap'>$Lang::tr{'vpn delayed start'}: <img src='/blob.gif' alt='*' /><img src='/blob.gif' alt='*' /></td> @@ -2492,13 +2418,6 @@ print <<END <td ><input type='text' name='RW_NET' value='$cgiparams{'RW_NET'}' /></td> </tr> </table> -<p>$Lang::tr{'vpn watch'}:<input type='checkbox' name='VPN_WATCH' $checked{'VPN_WATCH'} /></p> -<p>PLUTO DEBUG = -crypt:<input type='checkbox' name='DBG_CRYPT' $checked{'DBG_CRYPT'} />,  -parsing:<input type='checkbox' name='DBG_PARSING' $checked{'DBG_PARSING'} />,  -emitting:<input type='checkbox' name='DBG_EMITTING' $checked{'DBG_EMITTING'} />,  -control:<input type='checkbox' name='DBG_CONTROL' $checked{'DBG_CONTROL'} />,  -dns:<input type='checkbox' name='DBG_DNS' $checked{'DBG_DNS'} />  <hr /> <table width='100%'> <tr> diff --git a/lfs/strongswan b/lfs/strongswan index d0d533d..3d22086 100644 --- a/lfs/strongswan +++ b/lfs/strongswan @@ -24,7 +24,7 @@ include Config -VER = 4.6.4 +VER = 5.0.0 THISAPP = strongswan-$(VER) DL_FILE = $(THISAPP).tar.bz2 @@ -32,6 +32,12 @@ DL_FROM = $(URL_IPFIRE) DIR_APP = $(DIR_SRC)/$(THISAPP) TARGET = $(DIR_INFO)/$(THISAPP) +ifeq "$(MACHINE)" "i586" + PADLOCK = --enable-padlock +else + PADLOCK = --disable-padlock +endif + ############################################################################### # Top-level Rules ############################################################################### @@ -40,7 +46,7 @@ objects = $(DL_FILE) $(DL_FILE) = $(DL_FROM)/$(DL_FILE) -$(DL_FILE)_MD5 = 4c0999c42faa0860ae0afc4f8efd9d04 +$(DL_FILE)_MD5 = c8b861305def7c0abae04f7bbefec212 install : $(TARGET) @@ -73,18 +79,19 @@ $(TARGET) : $(patsubst %,$(DIR_DL)/%,$(objects)) cd $(DIR_APP) && patch -Np1 -i $(DIR_SRC)/src/patches/strongswan-4.5.3_ipfire.patch - # Customize the welcome banner. - sed -i $(DIR_APP)/src/pluto/modecfg.c \ - -e 's/^#define.*DEFAULT_UNITY_BANNER.*/#define DEFAULT_UNITY_BANNER "Welcome to IPFire - An Open Source Firewall Solution.\\n"/' - - cd $(DIR_APP) && ./configure --prefix="/usr" --sysconfdir="/etc" \ - --enable-cisco-quirks \ - --enable-curl \ - --enable-nat-transport + cd $(DIR_APP) && ./configure \ + --prefix="/usr" \ + --sysconfdir="/etc" \ + --enable-curl \ + --enable-openssl \ + $(PADLOCK) cd $(DIR_APP) && make $(MAKETUNING) cd $(DIR_APP) && make install + # Remove all library files we don't want or need. + rm -vf /usr/lib/ipsec/plugins/*.{,l}a + -rm -rfv /etc/rc*.d/*ipsec cd $(DIR_SRC) && cp src/initscripts/init.d/ipsec /etc/rc.d/init.d/ipsec rm -f /etc/ipsec.conf /etc/ipsec.secrets diff --git a/src/initscripts/init.d/tmpfs b/src/initscripts/init.d/tmpfs index 0e5a1e1..848dec6 100644 --- a/src/initscripts/init.d/tmpfs +++ b/src/initscripts/init.d/tmpfs @@ -43,10 +43,6 @@ case "$1" in mkdir -p /var/run/mysql chown mysql:mysql /var/run/mysql fi - if [ ! -e /var/run/pluto ]; then - mkdir -p /var/run/pluto - chmod 700 /var/run/pluto - fi if [ ! -e /var/run/saslauthd ]; then mkdir -p /var/run/saslauthd fi diff --git a/src/misc-progs/ipsecctrl.c b/src/misc-progs/ipsecctrl.c index 0b05177..65a96e0 100644 --- a/src/misc-progs/ipsecctrl.c +++ b/src/misc-progs/ipsecctrl.c @@ -78,7 +78,6 @@ void ipsec_norules() { safe_system("/sbin/iptables -F IPSECINPUT"); safe_system("/sbin/iptables -F IPSECFORWARD"); safe_system("/sbin/iptables -F IPSECOUTPUT"); - } /* @@ -87,8 +86,7 @@ void ipsec_norules() { int decode_line (char *s, char **key, char **name, - char **type, - char **interface + char **type ) { int count = 0; *key = NULL; @@ -108,8 +106,6 @@ int decode_line (char *s, *name = result; if (count == 4) *type = result; - if (count == 27) - *interface = result; count++; result = strsep(&s, ","); } @@ -128,11 +124,6 @@ int decode_line (char *s, return 0; } - if (! (strcmp(*interface, "RED") == 0 || strcmp(*interface, "GREEN") == 0 || - strcmp(*interface, "ORANGE") == 0 || strcmp(*interface, "BLUE") == 0)) { - fprintf(stderr, "Bad interface name: %s\n", *interface); - return 0; - } //it's a valid & active line return 1; } @@ -140,69 +131,48 @@ int decode_line (char *s, /* issue ipsec commmands to turn on connection 'name' */ -void turn_connection_on (char *name, char *type) { -/* - Rename the connection and run ipsec update and rename it back to readd - a deleted connection. Because ipsec update ignores connection that have - not changed since last load. -*/ +void turn_connection_on(char *name, char *type) { + /* + * To bring up a connection, we need to reload the configuration + * and issue ipsec up afterwards. To make sure the connection + * is not established from the start, we bring it down in advance. + */ char command[STRING_SIZE]; - memset(command, 0, STRING_SIZE); - snprintf(command, STRING_SIZE - 1, - "sed -i -e 's|^conn %s$|conn %s-renamed|g' /var/ipfire/vpn/ipsec.conf >/dev/null", name, name); - safe_system(command); - // Down and delete IKEv2 Tunnel before ipsec update + // Bring down the connection (if established). snprintf(command, STRING_SIZE - 1, - "/usr/sbin/ipsec stroke down %s >/dev/null", name); + "/usr/sbin/ipsec down %s >/dev/null", name); safe_system(command); - snprintf(command, STRING_SIZE - 1, - "/usr/sbin/ipsec stroke delete %s >/dev/null", name); - safe_system(command); - - safe_system("/etc/rc.d/init.d/ipsec update >/dev/null"); - sleep(1); + // Reload the configuration into the daemon. + safe_system("/usr/sbin/ipsec reload >/dev/null 2>&1"); - // Back to original name - snprintf(command, STRING_SIZE - 1, - "sed -i -e 's|^conn %s-renamed$|conn %s|g' /var/ipfire/vpn/ipsec.conf >/dev/null", name, name); - safe_system(command); - - // Down and delete IKEv2 Tunnel before ipsec update - snprintf(command, STRING_SIZE - 1, - "/usr/sbin/ipsec stroke down %s-renamed >/dev/null", name); - safe_system(command); - snprintf(command, STRING_SIZE - 1, - "/usr/sbin/ipsec stroke delete %s-renamed >/dev/null", name); - safe_system(command); - - safe_system("/etc/rc.d/init.d/ipsec update >/dev/null"); + // Bring the connection up again. + snprintf(command, STRING_SIZE - 1, + "/usr/sbin/ipsec up %s >/dev/null", name); + safe_system(command); } + /* issue ipsec commmands to turn off connection 'name' */ void turn_connection_off (char *name) { + /* + * To turn off a connection, all SAs must be turned down. + * After that, the configuration must be reloaded. + */ char command[STRING_SIZE]; - memset(command, 0, STRING_SIZE); - snprintf(command, STRING_SIZE - 1, - "/usr/sbin/ipsec whack --delete --name %s >/dev/null", name); - safe_system(command); - snprintf(command, STRING_SIZE - 1, - "/usr/sbin/ipsec stroke down %s >/dev/null", name); - safe_system(command); + + // Bring down the connection. snprintf(command, STRING_SIZE - 1, - "/usr/sbin/ipsec stroke delete %s >/dev/null", name); + "/usr/sbin/ipsec down %s >/dev/null", name); safe_system(command); - safe_system("/usr/sbin/ipsec whack --rereadall >/dev/null"); - safe_system("/usr/sbin/ipsec stroke rereadall >/dev/null"); - + // Reload, so the connection is dropped. + safe_system("/usr/sbin/ipsec reload >/dev/null 2>&1"); } - int main(int argc, char *argv[]) { - char configtype[STRING_SIZE]; char redtype[STRING_SIZE] = ""; struct keyvalue *kv = NULL; @@ -218,26 +188,15 @@ int main(int argc, char *argv[]) { if (strcmp(argv[1], "I") == 0) { - safe_system("/usr/sbin/ipsec whack --status"); - safe_system("/usr/sbin/ipsec stroke status"); + safe_system("/usr/sbin/ipsec status"); exit(0); } if (strcmp(argv[1], "R") == 0) { - safe_system("/usr/sbin/ipsec whack --rereadall >/dev/null"); - safe_system("/usr/sbin/ipsec stroke rereadall >/dev/null"); + safe_system("/usr/sbin/ipsec reload >/dev/null 2>&1"); exit(0); } - /* Get vpnwatch pid */ - - - if ((argc == 2) && (file = fopen("/var/run/vpn-watch.pid", "r"))) { - safe_system("kill -9 $(cat /var/run/vpn-watch.pid)"); - safe_system("unlink /var/run/vpn-watch.pid"); - close(file); - } - /* FIXME: workaround for pclose() issue - still no real idea why * this is happening */ signal(SIGCHLD, SIG_DFL); @@ -245,16 +204,10 @@ int main(int argc, char *argv[]) { /* handle operations that doesn't need start the ipsec system */ if (argc == 2) { if (strcmp(argv[1], "D") == 0) { - /* Only shutdown pluto if it really is running */ - /* Get pluto pid */ - if (file = fopen("/var/run/pluto.pid", "r")) { - safe_system("/etc/rc.d/init.d/ipsec stop 2> /dev/null >/dev/null"); - close(file); - } + safe_system("/usr/sbin/ipsec stop >/dev/null 2>&1"); ipsec_norules(); exit(0); } - } /* read vpn config */ @@ -300,97 +253,69 @@ int main(int argc, char *argv[]) { char if_blue[STRING_SIZE] = ""; char s[STRING_SIZE]; - if (!(file = fopen(CONFIG_ROOT "/vpn/config", "r"))) { - fprintf(stderr, "Couldn't open vpn settings file"); - exit(1); - } - while (fgets(s, STRING_SIZE, file) != NULL) { - char *key; - char *name; - char *type; - char *interface; - if (!decode_line(s,&key,&name,&type,&interface)) - continue; - /* search interface */ - if (!enable_red && strcmp (interface, "RED") == 0) { - // when RED is up, find interface name in special file - FILE *ifacefile = NULL; - if ((ifacefile = fopen(CONFIG_ROOT "/red/iface", "r"))) { - if (fgets(if_red, STRING_SIZE, ifacefile)) { - if (if_red[strlen(if_red) - 1] == '\n') - if_red[strlen(if_red) - 1] = '\0'; - } - fclose (ifacefile); - - if (VALID_DEVICE(if_red)) - enable_red+=2; // present and running - } - } - - if (!enable_green && strcmp (interface, "GREEN") == 0) { - enable_green = 1; - findkey(kv, "GREEN_DEV", if_green); - if (VALID_DEVICE(if_green)) - enable_green++; - else - fprintf(stderr, "IPSec enabled on green but green interface is invalid or not found\n"); + // when RED is up, find interface name in special file + FILE *ifacefile = NULL; + if ((ifacefile = fopen(CONFIG_ROOT "/red/iface", "r"))) { + if (fgets(if_red, STRING_SIZE, ifacefile)) { + if (if_red[strlen(if_red) - 1] == '\n') + if_red[strlen(if_red) - 1] = '\0'; } + fclose (ifacefile); - if (!enable_orange && strcmp (interface, "ORANGE") == 0) { - enable_orange = 1; - findkey(kv, "ORANGE_DEV", if_orange); - if (VALID_DEVICE(if_orange)) - enable_orange++; - else - fprintf(stderr, "IPSec enabled on orange but orange interface is invalid or not found\n"); - } + if (VALID_DEVICE(if_red)) + enable_red++; + } - if (!enable_blue && strcmp (interface, "BLUE") == 0) { - enable_blue++; - findkey(kv, "BLUE_DEV", if_blue); - if (VALID_DEVICE(if_blue)) - enable_blue++; - else - fprintf(stderr, "IPSec enabled on blue but blue interface is invalid or not found\n"); + // Check if GREEN is enabled. + findkey(kv, "GREEN_DEV", if_green); + if (VALID_DEVICE(if_green)) + enable_green++; + else + fprintf(stderr, "IPSec enabled on green but green interface is invalid or not found\n"); + + // Check if ORANGE is enabled. + findkey(kv, "ORANGE_DEV", if_orange); + if (VALID_DEVICE(if_orange)) + enable_orange++; + else + fprintf(stderr, "IPSec enabled on orange but orange interface is invalid or not found\n"); + + // Check if BLUE is enabled. + findkey(kv, "BLUE_DEV", if_blue); + if (VALID_DEVICE(if_blue)) + enable_blue++; + else + fprintf(stderr, "IPSec enabled on blue but blue interface is invalid or not found\n"); - } - } - fclose(file); freekeyvalues(kv); - // do nothing if something is in error condition - if ((enable_red==1) || (enable_green==1) || (enable_orange==1) || (enable_blue==1) ) - exit(1); - // exit if nothing to do - if ( (enable_red+enable_green+enable_orange+enable_blue) == 0 ) + if ((enable_red+enable_green+enable_orange+enable_blue) == 0) exit(0); // open needed ports - // todo: read a nat_t indicator to allow or not openning UDP/4500 - if (enable_red==2) + if (enable_red > 0) open_physical(if_red, 4500); - if (enable_green==2) + if (enable_green > 0) open_physical(if_green, 4500); - if (enable_orange==2) + if (enable_orange > 0) open_physical(if_orange, 4500); - if (enable_blue==2) + if (enable_blue > 0) open_physical(if_blue, 4500); // start the system if ((argc == 2) && strcmp(argv[1], "S") == 0) { - safe_system("/etc/rc.d/init.d/ipsec restart >/dev/null"); - safe_system("/usr/local/bin/vpn-watch &"); + safe_system("/usr/sbin/ipsec restart >/dev/null"); exit(0); } // it is a selective start or stop // second param is only a number 'key' if ((argc == 2) || strspn(argv[2], NUMBERS) != strlen(argv[2])) { - fprintf(stderr, "Bad arg\n"); + fprintf(stderr, "Bad arg: %s\n", argv[2]); usage(); exit(1); } @@ -404,26 +329,17 @@ int main(int argc, char *argv[]) { char *key; char *name; char *type; - char *interface; - if (!decode_line(s,&key,&name,&type,&interface)) + if (!decode_line(s,&key,&name,&type)) continue; - // start/stop a vpn if belonging to specified interface - if (strcmp(argv[1], interface) == 0 ) { - if (strcmp(argv[2], "0")==0) - turn_connection_off (name); - else - turn_connection_on (name, type); - continue; - } // is it the 'key' requested ? if (strcmp(argv[2], key) != 0) continue; + // Start or Delete this Connection if (strcmp(argv[1], "S") == 0) turn_connection_on (name, type); - else - if (strcmp(argv[1], "D") == 0) + else if (strcmp(argv[1], "D") == 0) turn_connection_off (name); else { fprintf(stderr, "Bad command\n"); @@ -431,5 +347,6 @@ int main(int argc, char *argv[]) { } } fclose(file); + return 0; } diff --git a/src/scripts/vpn-watch b/src/scripts/vpn-watch deleted file mode 100755 index 7eae873..0000000 --- a/src/scripts/vpn-watch +++ /dev/null @@ -1,83 +0,0 @@ -#!/usr/bin/perl -################################################## -##### VPN-Watch.pl Version 0.7 ##### -################################################## -# # -# VPN-Watch is part of the IPFire Firewall # -# # -################################################## - -use strict; - -require '/var/ipfire/general-functions.pl'; -my @vpnsettings; -my $i = 0; -my $file = "/var/run/vpn-watch.pid"; -my $debug = 0; - -if ( -e $file ){ - logger("There my be another vpn-watch runnning because $file exists, vpn-watch will try kill the process."); - open(FILE, "<$file"); - my $PID = <FILE>; - close(FILE); - system("kill -9 $PID"); - } - -system("echo $$ > $file"); -my $round=0; -while ( $i == 0){ - if ($debug){logger("We will wait 60 seconds before next action.");} - sleep(60); - - $round++; - - # Reset roundcounter after 10 min. To do established check. - if ($round > 9) { $round=0; } - - if (open(FILE, "<${General::swroot}/vpn/config")) { @vpnsettings = <FILE>; - close(FILE); - unless(@vpnsettings) {exit 1;} - } - -my $status = `ipsec status`; -foreach (@vpnsettings){ - my @settings = split(/,/,$_); - - chomp($settings[30]); - if ($settings[27] ne 'RED'){next;} - if ($settings[4] ne 'net'){next;} - if ($settings[1] ne 'on'){next;}chomp($settings[29]); - if ($settings[29] ne 'on'){next;} - - my $remotehostname = $settings[11]; - - if ($debug){logger("Checking connection to $remotehostname.");} - - my $remoteip = `/usr/bin/ping -c 1 $remotehostname 2>/dev/null | head -n1 | awk '{print \$3}' | tr -d '()' | tr -d ':'`;chomp($remoteip); - if ($remoteip eq ""){next;if ($debug){logger("Unable to resolve $remotehostname.");}} - my $ipmatch= `echo "$status" | grep '$remoteip' | grep '$settings[2]'`; - my $established= `echo "$status" | grep '$settings[2]' | grep -e 'erouted;' -e 'INSTALLED'`; - my $known= `echo "$status" | grep '$settings[2]'`; - - if ( $ipmatch eq '' && $known ne '' ){ - logger("Remote IP for host $remotehostname($remoteip) has changed, restarting ipsec."); - system("/usr/local/bin/ipsecctrl S $settings[0]"); - $round=0; - } - - if ($debug){logger("Round=".$round." and established=".$established);} - - if ( ($round == 0) && ($established eq '')) { - logger("Connection to $remotehostname($remoteip) not erouted, restarting ipsec."); - system("/usr/local/bin/ipsecctrl S $settings[0]"); - $round=0; - - } - } - if ($debug){logger("All connections may be fine nothing was done.");} -} - -sub logger { - my $log = shift; - system("logger -t vpnwatch \"$log\""); -} hooks/post-receive -- IPFire 2.x development tree

1 0

[git.ipfire.org] IPFire 3.x development tree branch, master, updated. a33f4021f466013d38bcae35b18bf8a07b50a250
by git＠ipfire.org 12 Aug '12

12 Aug '12

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "IPFire 3.x development tree". The branch, master has been updated via a33f4021f466013d38bcae35b18bf8a07b50a250 (commit) via 36c5c701c2fbd01760b36bcc7e3d2221b6bbf84b (commit) from 028e459b5a98c541f22cb4bb78d6064ca9d1c4c0 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit a33f4021f466013d38bcae35b18bf8a07b50a250 Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Aug 12 21:13:16 2012 +0000 rrdtool: Update to 1.4.7. Create an extra package for the python bindings. commit 36c5c701c2fbd01760b36bcc7e3d2221b6bbf84b Author: Michael Tremer <michael.tremer(a)ipfire.org> Date: Sun Aug 12 21:12:30 2012 +0000 openvswitch: Fix unit file. openvswitch got stuck when a switch has to be created at boot time because network.target was never reached. ----------------------------------------------------------------------- Summary of changes: openvswitch/openvswitch.nm | 2 +- openvswitch/systemd/openvswitch.service | 4 ++-- rrdtool/rrdtool.nm | 14 ++++++++++++-- 3 files changed, 15 insertions(+), 5 deletions(-) Difference in files: diff --git a/openvswitch/openvswitch.nm b/openvswitch/openvswitch.nm index 928d152..b4dfa3b 100644 --- a/openvswitch/openvswitch.nm +++ b/openvswitch/openvswitch.nm @@ -5,7 +5,7 @@ name = openvswitch version = 1.7.0 -release = 2 +release = 3 groups = Networking/Tools url = http://openvswitch.org/ diff --git a/openvswitch/systemd/openvswitch.service b/openvswitch/systemd/openvswitch.service index 7aeae19..6186a8f 100644 --- a/openvswitch/systemd/openvswitch.service +++ b/openvswitch/systemd/openvswitch.service @@ -1,11 +1,11 @@ [Unit] Description=Open vSwitch -After=syslog.target network.target +After=syslog.target [Service] Type=oneshot ExecStartPre=/usr/share/openvswitch/scripts/ovs-ctl enable-protocol --protocol=gre -ExecStart=/usr/share/openvswitch/scripts/ovs-ctl --system-id=random start +ExecStart=/usr/share/openvswitch/scripts/ovs-ctl --system-id=random start --delete-bridges ExecStop=/usr/share/openvswitch/scripts/ovs-ctl stop RemainAfterExit=yes diff --git a/rrdtool/rrdtool.nm b/rrdtool/rrdtool.nm index cbfb3d4..fbf994b 100644 --- a/rrdtool/rrdtool.nm +++ b/rrdtool/rrdtool.nm @@ -4,7 +4,7 @@ ############################################################################### name = rrdtool -version = 1.4.5 +version = 1.4.7 release = 1 groups = Applications/Databases @@ -60,12 +60,22 @@ end packages package %{name} - end package %{name}-libs template LIBS end + package python-%{name} + summary = Python bindings for rrdtool %{thisver}. + description = %{summary} + + requires = %{name}-libs = %{thisver} + + files + /usr/lib*/python* + end + end + package %{name}-devel template DEVEL end hooks/post-receive -- IPFire 3.x development tree

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

IPFire-SCM August 2012