aboutsummaryrefslogtreecommitdiffstats
path: root/include/net
Commit message (Collapse)AuthorAge
...
* | | [PKT_SCHED]: Pass real namespace in net scheduler classifiers.Denis V. Lunev2008-03-27
|/ / | | | | | | | | Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][IPV6] flowlabels - make flowlabels per namespaceBenjamin Thery2008-03-26
| | | | | | | | | | | | | | | | | | This patch introduces a new member, fl_net, in struct ip6_flowlabel. This allows to create labels with the same value in different namespaces. Signed-off-by: Benjamin Thery <benjamin.thery@bull.net> Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][IPV6] anycast - handle several network namespaceDaniel Lezcano2008-03-26
| | | | | | | | | | | | | | | | | | Make use of the network namespace information to have this protocol to handle several network namespace. Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: Benjamin Thery <benjamin.thery@bull.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][ICMP]: Make ctl tables for ICMP sysctls per-net.Pavel Emelyanov2008-03-26
| | | | | | | | | | | | | | | | | | | | | | | | Add some flesh to ipv4_sysctl_init_net and ipv4_sysctl_exit_net, i.e. copy the table, alter .data pointers and register it per-net. Other ipv4_table's sysctls are now global, but this is going to change once sysctl permissions patches migrate from -mm tree to mainline in 2.6.26 merge window :) Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][ICMP]: Move ICMP sysctls on struct net.Pavel Emelyanov2008-03-26
| | | | | | | | | | | | | | | | Initialization is moved to icmp_sk_init, all the places, that refer to them use init_net for now. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS]: Compilation warnings under CONFIG_NET_NS.Denis V. Lunev2008-03-26
| | | | | | | | | | | | | | | | | | | | | | | | | | Recent commits from YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org> have been introduced a several compilation warnings 'assignment discards qualifiers from pointer target type' due to extra const modifier in the inline call parameters of {dev|sock|twsk}_net_set. Drop it. Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETFILTER]: nf_conntrack_sip: support multiple media channelsPatrick McHardy2008-03-25
| | | | | | | | | | | | | | | | Add support for multiple media channels and use it to create expectations for video streams when present. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETFILTER]: nf_conntrack_sip: create signalling expectationsPatrick McHardy2008-03-25
| | | | | | | | | | | | | | | | | | | | Create expectations for incoming signalling connections when seeing a REGISTER request. This is needed when the registrar uses a different source port number for signalling messages and for receiving incoming calls from other endpoints than the registrar. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETFILTER]: Add nf_inet_addr_cmp()Patrick McHardy2008-03-25
| | | | | | | | | | Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETFILTER]: nf_conntrack: introduce expectation classes and policiesPatrick McHardy2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Introduce expectation classes and policies. An expectation class is used to distinguish different types of expectations by the same helper (for example audio/video/t.120). The expectation policy is used to hold the maximum number of expectations and the initial timeout for each class. The individual classes are isolated from each other, which means that for example an audio expectation will only evict other audio expectations. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETFILTER]: nf_conntrack_expect: support inactive expectationsPatrick McHardy2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | This is useful for the SIP helper and signalling expectations. We don't want to create a full-blown expectation with a wildcard as source based on a single UDP packet, but need to know the final port anyways. With inactive expectations we can register the expectation and reserve the tuple, but wait for confirmation from the registrar before activating it. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETFILTER]: nf_conntrack_expect: constify nf_ct_expect_init argumentsPatrick McHardy2008-03-25
| | | | | | | | | | Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETFILTER]: nf_conntrack: fix NF_CT_TUPLE_DUMP for IPv4Patrick McHardy2008-03-25
| | | | | | | | | | | | | | | | | | | | NF_CT_TUPLE_DUMP prints IPv4 addresses as IPv6, fix this and use printk (guarded by #ifdef DEBUG) directly instead of pr_debug since the tuple is usually printed at the end of line and we don't want to include a log-level. Signed-off-by: Patrick McHardy <kaber@trash.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge branch 'net-2.6.26-netns-20080326' of ↵David S. Miller2008-03-25
|\ \ | | | | | | | | | git://git.linux-ipv6.org/gitroot/yoshfuji/linux-2.6-dev
| * | [NET] NETNS: Omit namespace comparision without CONFIG_NET_NS.YOSHIFUJI Hideaki2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Introduce an inline net_eq() to compare two namespaces. Without CONFIG_NET_NS, since no namespace other than &init_net exists, it is always 1. We do not need to convert 1) inline vs inline and 2) inline vs &init_net comparisons. Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
| * | [NET] NETNS: Omit neigh_parms->net and pneigh_entry->net without CONFIG_NET_NS.YOSHIFUJI Hideaki2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | Introduce neigh_parms/pneigh_entry inlines: neigh_parms_net(), pneigh_net(). Without CONFIG_NET_NS, no namespace other than &init_net exists. Let's explicitly define them to help compiler optimizations. Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
| * | [NET] NETNS: Omit sock->sk_net without CONFIG_NET_NS.YOSHIFUJI Hideaki2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | Introduce per-sock inlines: sock_net(), sock_net_set() and per-inet_timewait_sock inlines: twsk_net(), twsk_net_set(). Without CONFIG_NET_NS, no namespace other than &init_net exists. Let's explicitly define them to help compiler optimizations. Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
* | | Merge branch 'master' of ↵David S. Miller2008-03-25
|\ \ \ | |/ / |/| | | | | git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6.26
| * | wireless: add wiphy channel freq to channel struct lookup helperJohannes Berg2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | Add ieee80211_get_channel() which gets you a channel struct for a specific wiphy if that channel is present in that wiphy. Signed-off-by: Johannes Berg <johannes@sipsolutions.net> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * | mac80211: allows driver to request a Phase 1 RX keyEmmanuel Grumbach2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch makes mac80211 able to send a phase1 key for TKIP decryption. This is needed for drivers that don't do the rekeying by themselves (i.e. iwlwifi). Upon IV16 wrap around, the packet is decrypted in SW, if decryption is ok, mac80211 calls to update_tkip_key with a new phase 1 RX key. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
| * | mac80211: get a TKIP phase key from skbEmmanuel Grumbach2008-03-25
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch makes mac80211 able to compute a TKIP key from an skb. The requested key can be a phase 1 or a phase 2 key. This is useful for drivers who need to provide tkip key to their HW to enable HW encryption. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Tomas Winkler <tomas.winkler@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
* | | [IPV6]: Support Source Address Selection API (RFC5014).YOSHIFUJI Hideaki2008-03-24
| | | | | | | | | | | | Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
* | | [IPV6]: Optimize hop-limit determination.YOSHIFUJI Hideaki2008-03-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Last part of hop-limit determination is always: hoplimit = dst_metric(dst, RTAX_HOPLIMIT); if (hoplimit < 0) hoplimit = ipv6_get_hoplimit(dst->dev). Let's consolidate it as ip6_dst_hoplimit(dst). Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
* | | [IPV4,IPV6]: Share cork.rt between IPv4 and IPv6.YOSHIFUJI Hideaki2008-03-24
| | | | | | | | | | | | Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
* | | [XFRM] MIP6: Fix address keys for routing search.YOSHIFUJI Hideaki2008-03-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Each MIPv6 XFRM state (DSTOPT/RH2) holds either destination or source address to be mangled in the IPv6 header (that is "CoA"). On Inter-MN communication after both nodes binds each other, they use route optimized traffic two MIPv6 states applied, and both source and destination address in the IPv6 header are replaced by the states respectively. The packet format is correct, however, next-hop routing search are not. This patch fixes it by remembering address pairs for later states. Based on patch from Masahide NAKAMURA <nakam@linux-ipv6.org>. Signed-off-by: Masahide NAKAMURA <nakam@linux-ipv6.org> Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
* | | [NETNS]: Drop packets in the non-initial namespace on the per/protocol basis.Denis V. Lunev2008-03-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | IP layer now can handle multiple namespaces normally. So, process such packets normally and drop them only if the transport layer is not aware about namespaces. Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | [NETNS]: Add namespace parameter to ip_cmsg_send.Denis V. Lunev2008-03-24
| | | | | | | | | | | | | | | | | | | | | Pass the init_net there for now. Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | [NETNS]: Add namespace parameter to ip_options_get(...).Denis V. Lunev2008-03-24
| | | | | | | | | | | | | | | | | | | | | Pass the init_net there for now. Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | | [NETNS]: Add namespace parameter to ip_options_compile.Denis V. Lunev2008-03-24
|/ / | | | | | | | | | | | | | | | | ip_options_compile uses inet_addr_type which requires a namespace. The packet argument is optional, so parameter is the only way to obtain it. Pass the init_net there for now. Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge branch 'master' of ../net-2.6/David S. Miller2008-03-24
|\| | | | | | | | | | | Conflicts: net/ipv6/ndisc.c
| * [SCTP]: Fix build warnings with IPV6 disabled.David S. Miller2008-03-21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Introduced by 270637abff0cdf848b910b9f96ad342e1da61c66 ("[SCTP]: Fix a race between module load and protosw access") Reported by Gabriel C: In file included from net/sctp/sm_statetable.c:50: include/net/sctp/sctp.h: In function 'sctp_v6_pf_init': include/net/sctp/sctp.h:392: warning: 'return' with a value, in function returning void In file included from net/sctp/sm_statefuns.c:62: include/net/sctp/sctp.h: In function 'sctp_v6_pf_init': include/net/sctp/sctp.h:392: warning: 'return' with a value, in function returning void ... Signed-off-by: David S. Miller <davem@davemloft.net>
* | [SCTP]: Remove redundant wrapper functions.Florian Westphal2008-03-24
| | | | | | | | | | | | | | | | | | | | sctp_datamsg_free and sctp_datamsg_track are just aliases for sctp_datamsg_put and sctp_chunk_hold, respectively. Saves 32 Bytes on x86. Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [TCP]: Shrink syncookie_secret by 8 byte.Florian Westphal2008-03-24
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | the first u32 copied from syncookie_secret is overwritten by the minute-counter four lines below. After adjusting the destination address, the size of syncookie_secret can be reduced accordingly. AFAICS, the only other user of syncookie_secret[] is the ipv6 syncookie support. Because ipv6 syncookies only grab 44 bytes from syncookie_secret[], this shouldn't affect them in any way. With fixes from Glenn Griffin. Signed-off-by: Florian Westphal <fw@strlen.de> Acked-by: Glenn Griffin <ggriffin.kernel@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NET]: include/net/route.h - remove duplicate includeJoe Perches2008-03-24
| | | | | | | | | | Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [RAW]: Add raw_hashinfo member on struct proto.Pavel Emelyanov2008-03-22
| | | | | | | | | | | | | | | | | | | | | | | | Sorry for the patch sequence confusion :| but I found that the similar thing can be done for raw sockets easily too late. Expand the proto.h union with the raw_hashinfo member and use it in raw_prot and rawv6_prot. This allows to drop the protocol specific versions of hash and unhash callbacks. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [UDP]: Make full use of proto.h.udp_hash innovation.Pavel Emelyanov2008-03-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | After this we have only udp_lib_get_port to get the port and two stubs for ipv4 and ipv6. No difference in udp and udplite except for initialized h.udp_hash member. I tried to find a graceful way to drop the only difference between udp_v4_get_port and udp_v6_get_port (i.e. the rcv_saddr comparison routine), but adding one more callback on the struct proto didn't appear such :( Maybe later. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [SOCK]: Add udp_hash member to struct proto.Pavel Emelyanov2008-03-22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Inspired by the commit ab1e0a13 ([SOCK] proto: Add hashinfo member to struct proto) from Arnaldo, I made similar thing for UDP/-Lite IPv4 and -v6 protocols. The result is not that exciting, but it removes some levels of indirection in udpxxx_get_port and saves some space in code and text. The first step is to union existing hashinfo and new udp_hash on the struct proto and give a name to this union, since future initialization of tcpxxx_prot, dccp_vx_protinfo and udpxxx_protinfo will cause gcc warning about inability to initialize anonymous member this way. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [IPV4]: Remove unused ip_options->is_data.Denis V. Lunev2008-03-22
| | | | | | | | | | | | | | | | | | ip_options->is_data is assigned only and never checked. The structure is not a part of kernel interface to the userspace. So, it is safe to remove this field. Signed-off-by: Denis V. Lunev <den@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [TCP]: TCP_DEFER_ACCEPT updates - process as establishedPatrick McManus2008-03-21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Change TCP_DEFER_ACCEPT implementation so that it transitions a connection to ESTABLISHED after handshake is complete instead of leaving it in SYN-RECV until some data arrvies. Place connection in accept queue when first data packet arrives from slow path. Benefits: - established connection is now reset if it never makes it to the accept queue - diagnostic state of established matches with the packet traces showing completed handshake - TCP_DEFER_ACCEPT timeouts are expressed in seconds and can now be enforced with reasonable accuracy instead of rounding up to next exponential back-off of syn-ack retry. Signed-off-by: Patrick McManus <mcmanus@ducksong.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | socket: SOCK_DEBUG type checkingStephen Hemminger2008-03-21
| | | | | | | | | | | | | | | | Use the inline trick (same as pr_debug) to get checking of debug statements even if no code is generated. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][IPV6] tcp6 - make proc per namespaceDaniel Lezcano2008-03-21
| | | | | | | | | | | | | | Make the proc for tcp6 to be per namespace. Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][IPV6] udp6 - make proc per namespaceDaniel Lezcano2008-03-21
| | | | | | | | | | | | | | | | The proc init/exit functions take a new network namespace parameter in order to register/unregister /proc/net/udp6 for a namespace. Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][IPV4] tcp - make proc handle the network namespacesDaniel Lezcano2008-03-21
| | | | | | | | | | | | | | | | This patch, like udp proc, makes the proc functions to take care of which namespace the socket belongs. Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NETNS][IPV6] udp - make proc handle the network namespaceDaniel Lezcano2008-03-21
| | | | | | | | | | | | | | | | This patch makes the common udp proc functions to take care of which socket they should show taking into account the namespace it belongs. Signed-off-by: Daniel Lezcano <dlezcano@fr.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | [NET]: Add per-connection option to set max TSO frame sizePeter P Waskiewicz Jr2008-03-21
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Update: My mailer ate one of Jarek's feedback mails... Fixed the parameter in netif_set_gso_max_size() to be u32, not u16. Fixed the whitespace issue due to a patch import botch. Changed the types from u32 to unsigned int to be more consistent with other variables in the area. Also brought the patch up to the latest net-2.6.26 tree. Update: Made gso_max_size container 32 bits, not 16. Moved the location of gso_max_size within netdev to be less hotpath. Made more consistent names between the sock and netdev layers, and added a define for the max GSO size. Update: Respun for net-2.6.26 tree. Update: changed max_gso_frame_size and sk_gso_max_size from signed to unsigned - thanks Stephen! This patch adds the ability for device drivers to control the size of the TSO frames being sent to them, per TCP connection. By setting the netdevice's gso_max_size value, the socket layer will set the GSO frame size based on that value. This will propogate into the TCP layer, and send TSO's of that size to the hardware. This can be desirable to help tune the bursty nature of TSO on a per-adapter basis, where one may have 1 GbE and 10 GbE devices coexisting in a system, one running multiqueue and the other not, etc. This can also be desirable for devices that cannot support full 64 KB TSO's, but still want to benefit from some level of segmentation offloading. Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge branch 'master' of ↵David S. Miller2008-03-21
|\| | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
| * [SCTP]: Fix a race between module load and protosw accessVlad Yasevich2008-03-20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There is a race is SCTP between the loading of the module and the access by the socket layer to the protocol functions. In particular, a list of addresss that SCTP maintains is not initialized prior to the registration with the protosw. Thus it is possible for a user application to gain access to SCTP functions before everything has been initialized. The problem shows up as odd crashes during connection initializtion when we try to access the SCTP address list. The solution is to refactor how we do registration and initialize the lists prior to registering with the protosw. Care must be taken since the address list initialization depends on some other pieces of SCTP initialization. Also the clean-up in case of failure now also needs to be refactored. Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com> Acked-by: Sridhar Samudrala <sri@us.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Merge branch 'master' of ↵David S. Miller2008-03-18
|\| | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6 Conflicts: drivers/net/wireless/rt2x00/rt2x00dev.c net/8021q/vlan_dev.c
| * xfrm: ->eth_proto is __be16Al Viro2008-03-18
| | | | | | | | | | Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
| * [NET]: Fix tbench regression in 2.6.25-rc1Zhang Yanmin2008-03-13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Comparing with kernel 2.6.24, tbench result has regression with 2.6.25-rc1. 1) On 2 quad-core processor stoakley: 4%. 2) On 4 quad-core processor tigerton: more than 30%. bisect located below patch. b4ce92775c2e7ff9cf79cca4e0a19c8c5fd6287b is first bad commit commit b4ce92775c2e7ff9cf79cca4e0a19c8c5fd6287b Author: Herbert Xu <herbert@gondor.apana.org.au> Date: Tue Nov 13 21:33:32 2007 -0800 [IPV6]: Move nfheader_len into rt6_info The dst member nfheader_len is only used by IPv6. It's also currently creating a rather ugly alignment hole in struct dst. Therefore this patch moves it from there into struct rt6_info. Above patch changes the cache line alignment, especially member __refcnt. I did a testing by adding 2 unsigned long pading before lastuse, so the 3 members, lastuse/__refcnt/__use, are moved to next cache line. The performance is recovered. I created a patch to rearrange the members in struct dst_entry. With Eric and Valdis Kletnieks's suggestion, I made finer arrangement. 1) Move tclassid under ops in case CONFIG_NET_CLS_ROUTE=y. So sizeof(dst_entry)=200 no matter if CONFIG_NET_CLS_ROUTE=y/n. I tested many patches on my 16-core tigerton by moving tclassid to different place. It looks like tclassid could also have impact on performance. If moving tclassid before metrics, or just don't move tclassid, the performance isn't good. So I move it behind metrics. 2) Add comments before __refcnt. On 16-core tigerton: If CONFIG_NET_CLS_ROUTE=y, the result with below patch is about 18% better than the one without the patch; If CONFIG_NET_CLS_ROUTE=n, the result with below patch is about 30% better than the one without the patch. With 32bit 2.6.25-rc1 on 8-core stoakley, the new patch doesn't introduce regression. Thank Eric, Valdis, and David! Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com> Acked-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
generated by cgit v1.2.2 (git 2.25.0) at 2025-03-22 07:41:27 -0400