Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6.25
Linus Torvalds [Tue, 29 Jan 2008 11:54:01 +0000 (22:54 +1100)]
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6.25: (1470 commits)
  [IPV6] ADDRLABEL: Fix double free on label deletion.
  [PPP]: Sparse warning fixes.
  [IPV4] fib_trie: remove unneeded NULL check
  [IPV4] fib_trie: More whitespace cleanup.
  [NET_SCHED]: Use nla_policy for attribute validation in ematches
  [NET_SCHED]: Use nla_policy for attribute validation in actions
  [NET_SCHED]: Use nla_policy for attribute validation in classifiers
  [NET_SCHED]: Use nla_policy for attribute validation in packet schedulers
  [NET_SCHED]: sch_api: introduce constant for rate table size
  [NET_SCHED]: Use typeful attribute parsing helpers
  [NET_SCHED]: Use typeful attribute construction helpers
  [NET_SCHED]: Use NLA_PUT_STRING for string dumping
  [NET_SCHED]: Use nla_nest_start/nla_nest_end
  [NET_SCHED]: Propagate nla_parse return value
  [NET_SCHED]: act_api: use PTR_ERR in tcf_action_init/tcf_action_get
  [NET_SCHED]: act_api: use nlmsg_parse
  [NET_SCHED]: act_api: fix netlink API conversion bug
  [NET_SCHED]: sch_netem: use nla_parse_nested_compat
  [NET_SCHED]: sch_atm: fix format string warning
  [NETNS]: Add namespace for ICMP replying code.
  ...

1113 files changed:
CREDITS
Documentation/feature-removal-schedule.txt
Documentation/networking/00-INDEX
Documentation/networking/bonding.txt
Documentation/networking/can.txt [new file with mode: 0644]
Documentation/networking/dccp.txt
Documentation/networking/ip-sysctl.txt
Documentation/networking/shaper.txt [deleted file]
Documentation/networking/udplite.txt
Documentation/networking/xfrm_proc.txt [new file with mode: 0644]
MAINTAINERS
arch/ia64/hp/sim/simeth.c
arch/powerpc/platforms/pasemi/Makefile
arch/powerpc/platforms/pasemi/dma_lib.c [new file with mode: 0644]
arch/powerpc/platforms/pasemi/pasemi.h
arch/ppc/8260_io/enet.c
arch/ppc/8260_io/fcc_enet.c
drivers/atm/ambassador.c
drivers/atm/he.c
drivers/connector/cn_queue.c
drivers/connector/connector.c
drivers/infiniband/core/addr.c
drivers/infiniband/core/cma.c
drivers/infiniband/hw/cxgb3/iwch_cm.c
drivers/net/3c501.c
drivers/net/3c507.c
drivers/net/3c515.c
drivers/net/7990.c
drivers/net/Kconfig
drivers/net/Makefile
drivers/net/a2065.c
drivers/net/amd8111e.c
drivers/net/at1700.c
drivers/net/b44.c
drivers/net/bnx2.c
drivers/net/bnx2.h
drivers/net/bnx2_fw.h
drivers/net/bnx2_fw2.h
drivers/net/bnx2x.c [new file with mode: 0644]
drivers/net/bnx2x.h [new file with mode: 0644]
drivers/net/bnx2x_fw_defs.h [new file with mode: 0644]
drivers/net/bnx2x_hsi.h [new file with mode: 0644]
drivers/net/bnx2x_init.h [new file with mode: 0644]
drivers/net/bnx2x_init_values.h [new file with mode: 0644]
drivers/net/bnx2x_reg.h [new file with mode: 0644]
drivers/net/bonding/bond_main.c
drivers/net/can/Kconfig [new file with mode: 0644]
drivers/net/can/Makefile [new file with mode: 0644]
drivers/net/can/vcan.c [new file with mode: 0644]
drivers/net/cassini.c
drivers/net/chelsio/common.h
drivers/net/chelsio/cxgb2.c
drivers/net/chelsio/espi.c
drivers/net/chelsio/espi.h
drivers/net/chelsio/sge.c
drivers/net/chelsio/sge.h
drivers/net/chelsio/subr.c
drivers/net/cpmac.c
drivers/net/cxgb3/adapter.h
drivers/net/cxgb3/common.h
drivers/net/cxgb3/cxgb3_main.c
drivers/net/cxgb3/cxgb3_offload.c
drivers/net/cxgb3/firmware_exports.h
drivers/net/cxgb3/l2t.c
drivers/net/cxgb3/regs.h
drivers/net/cxgb3/sge.c
drivers/net/cxgb3/t3_hw.c
drivers/net/cxgb3/version.h
drivers/net/cxgb3/xgmac.c
drivers/net/declance.c
drivers/net/e100.c
drivers/net/e1000/e1000_ethtool.c
drivers/net/e1000/e1000_hw.h
drivers/net/e1000/e1000_main.c
drivers/net/e1000e/82571.c
drivers/net/e1000e/defines.h
drivers/net/e1000e/e1000.h
drivers/net/e1000e/ethtool.c
drivers/net/e1000e/hw.h
drivers/net/e1000e/lib.c
drivers/net/e1000e/netdev.c
drivers/net/e1000e/param.c
drivers/net/e1000e/phy.c
drivers/net/eepro100.c
drivers/net/eexpress.c
drivers/net/enc28j60.c [new file with mode: 0644]
drivers/net/enc28j60_hw.h [new file with mode: 0644]
drivers/net/forcedeth.c
drivers/net/gianfar_sysfs.c
drivers/net/hamradio/6pack.c
drivers/net/hamradio/mkiss.c
drivers/net/hamradio/scc.c
drivers/net/hp100.c
drivers/net/ibm_newemac/core.c
drivers/net/ibmlana.c
drivers/net/igb/Makefile [new file with mode: 0644]
drivers/net/igb/e1000_82575.c [new file with mode: 0644]
drivers/net/igb/e1000_82575.h [new file with mode: 0644]
drivers/net/igb/e1000_defines.h [new file with mode: 0644]
drivers/net/igb/e1000_hw.h [new file with mode: 0644]
drivers/net/igb/e1000_mac.c [new file with mode: 0644]
drivers/net/igb/e1000_mac.h [new file with mode: 0644]
drivers/net/igb/e1000_nvm.c [new file with mode: 0644]
drivers/net/igb/e1000_nvm.h [new file with mode: 0644]
drivers/net/igb/e1000_phy.c [new file with mode: 0644]
drivers/net/igb/e1000_phy.h [new file with mode: 0644]
drivers/net/igb/e1000_regs.h [new file with mode: 0644]
drivers/net/igb/igb.h [new file with mode: 0644]
drivers/net/igb/igb_ethtool.c [new file with mode: 0644]
drivers/net/igb/igb_main.c [new file with mode: 0644]
drivers/net/ipg.c
drivers/net/ipg.h
drivers/net/irda/Kconfig
drivers/net/irda/Makefile
drivers/net/irda/act200l.c [deleted file]
drivers/net/irda/actisys.c [deleted file]
drivers/net/irda/ep7211_ir.c [deleted file]
drivers/net/irda/esi.c [deleted file]
drivers/net/irda/girbil.c [deleted file]
drivers/net/irda/irport.c [deleted file]
drivers/net/irda/irport.h [deleted file]
drivers/net/irda/litelink.c [deleted file]
drivers/net/irda/ma600.c [deleted file]
drivers/net/irda/mcp2120.c [deleted file]
drivers/net/irda/old_belkin.c [deleted file]
drivers/net/irda/smsc-ircc2.c
drivers/net/irda/tekram.c [deleted file]
drivers/net/irda/via-ircc.c
drivers/net/ixgb/ixgb_hw.c
drivers/net/ixgb/ixgb_hw.h
drivers/net/ixgb/ixgb_ids.h
drivers/net/ixgb/ixgb_main.c
drivers/net/ixgbe/ixgbe.h
drivers/net/ixgbe/ixgbe_82598.c
drivers/net/ixgbe/ixgbe_common.c
drivers/net/ixgbe/ixgbe_ethtool.c
drivers/net/ixgbe/ixgbe_main.c
drivers/net/ixgbe/ixgbe_phy.h
drivers/net/ixgbe/ixgbe_type.h
drivers/net/loopback.c
drivers/net/lp486e.c
drivers/net/mac89x0.c
drivers/net/mace.c
drivers/net/macvlan.c
drivers/net/myri10ge/myri10ge.c
drivers/net/netconsole.c
drivers/net/netxen/netxen_nic.h
drivers/net/netxen/netxen_nic_ethtool.c
drivers/net/netxen/netxen_nic_hw.c
drivers/net/netxen/netxen_nic_hw.h
drivers/net/netxen/netxen_nic_init.c
drivers/net/netxen/netxen_nic_isr.c
drivers/net/netxen/netxen_nic_main.c
drivers/net/netxen/netxen_nic_niu.c
drivers/net/netxen/netxen_nic_phan_reg.h
drivers/net/niu.c
drivers/net/ns83820.c
drivers/net/pasemi_mac.c
drivers/net/pasemi_mac.h
drivers/net/pcmcia/axnet_cs.c
drivers/net/pcmcia/fmvj18x_cs.c
drivers/net/pcmcia/pcnet_cs.c
drivers/net/pcnet32.c
drivers/net/plip.c
drivers/net/ppp_deflate.c
drivers/net/ppp_generic.c
drivers/net/ppp_synctty.c
drivers/net/pppoe.c
drivers/net/pppol2tp.c
drivers/net/qla3xxx.c
drivers/net/qla3xxx.h
drivers/net/r6040.c [new file with mode: 0644]
drivers/net/rrunner.c
drivers/net/s2io-regs.h
drivers/net/s2io.c
drivers/net/s2io.h
drivers/net/sgiseeq.c
drivers/net/shaper.c [deleted file]
drivers/net/sis900.c
drivers/net/sk98lin/skgemib.c
drivers/net/sk98lin/skgepnmi.c
drivers/net/sk98lin/skgesirq.c
drivers/net/skfp/hwmtm.c
drivers/net/skfp/smt.c
drivers/net/sky2.c
drivers/net/sky2.h
drivers/net/slhc.c
drivers/net/slip.c
drivers/net/smc9194.c
drivers/net/sundance.c
drivers/net/sungem.c
drivers/net/sungem.h
drivers/net/sunhme.c
drivers/net/sunhme.h
drivers/net/sunvnet.c
drivers/net/tehuti.c
drivers/net/tg3.c
drivers/net/tg3.h
drivers/net/tokenring/olympic.c
drivers/net/tokenring/olympic.h
drivers/net/tokenring/smctr.c
drivers/net/tulip/de4x5.c
drivers/net/tun.c
drivers/net/ucc_geth.c
drivers/net/usb/dm9601.c
drivers/net/via-rhine.c
drivers/net/wan/farsync.c
drivers/net/wan/hdlc_ppp.c
drivers/net/wan/hdlc_raw_eth.c
drivers/net/wan/lmc/lmc_main.c
drivers/net/wan/lmc/lmc_media.c
drivers/net/wan/lmc/lmc_proto.c
drivers/net/wan/lmc/lmc_proto.h
drivers/net/wan/pc300_drv.c
drivers/net/wan/pc300_tty.c
drivers/net/wan/sbni.c
drivers/net/wan/sdla.c
drivers/net/wan/wanxl.c
drivers/net/wireless/Kconfig
drivers/net/wireless/Makefile
drivers/net/wireless/adm8211.c
drivers/net/wireless/airo.c
drivers/net/wireless/ath5k/Makefile [new file with mode: 0644]
drivers/net/wireless/ath5k/ath5k.h [new file with mode: 0644]
drivers/net/wireless/ath5k/base.c [new file with mode: 0644]
drivers/net/wireless/ath5k/base.h [new file with mode: 0644]
drivers/net/wireless/ath5k/debug.c [new file with mode: 0644]
drivers/net/wireless/ath5k/debug.h [new file with mode: 0644]
drivers/net/wireless/ath5k/hw.c [new file with mode: 0644]
drivers/net/wireless/ath5k/hw.h [new file with mode: 0644]
drivers/net/wireless/ath5k/initvals.c [new file with mode: 0644]
drivers/net/wireless/ath5k/phy.c [new file with mode: 0644]
drivers/net/wireless/ath5k/reg.h [new file with mode: 0644]
drivers/net/wireless/ath5k/regdom.c [new file with mode: 0644]
drivers/net/wireless/ath5k/regdom.h [new file with mode: 0644]
drivers/net/wireless/atmel.c
drivers/net/wireless/b43/Kconfig
drivers/net/wireless/b43/Makefile
drivers/net/wireless/b43/b43.h
drivers/net/wireless/b43/debugfs.c
drivers/net/wireless/b43/dma.c
drivers/net/wireless/b43/dma.h
drivers/net/wireless/b43/leds.c
drivers/net/wireless/b43/lo.c
drivers/net/wireless/b43/main.c
drivers/net/wireless/b43/main.h
drivers/net/wireless/b43/nphy.c [new file with mode: 0644]
drivers/net/wireless/b43/nphy.h [new file with mode: 0644]
drivers/net/wireless/b43/phy.c
drivers/net/wireless/b43/phy.h
drivers/net/wireless/b43/pio.c [deleted file]
drivers/net/wireless/b43/pio.h [deleted file]
drivers/net/wireless/b43/tables.c
drivers/net/wireless/b43/tables.h
drivers/net/wireless/b43/tables_nphy.c [new file with mode: 0644]
drivers/net/wireless/b43/tables_nphy.h [new file with mode: 0644]
drivers/net/wireless/b43/wa.c [new file with mode: 0644]
drivers/net/wireless/b43/wa.h [new file with mode: 0644]
drivers/net/wireless/b43/xmit.c
drivers/net/wireless/b43/xmit.h
drivers/net/wireless/b43legacy/Kconfig
drivers/net/wireless/b43legacy/Makefile
drivers/net/wireless/b43legacy/b43legacy.h
drivers/net/wireless/b43legacy/debugfs.c
drivers/net/wireless/b43legacy/ilt.c
drivers/net/wireless/b43legacy/leds.c
drivers/net/wireless/b43legacy/leds.h
drivers/net/wireless/b43legacy/main.c
drivers/net/wireless/b43legacy/main.h
drivers/net/wireless/b43legacy/phy.c
drivers/net/wireless/b43legacy/phy.h
drivers/net/wireless/b43legacy/radio.c
drivers/net/wireless/b43legacy/radio.h
drivers/net/wireless/b43legacy/rfkill.c [new file with mode: 0644]
drivers/net/wireless/b43legacy/rfkill.h [new file with mode: 0644]
drivers/net/wireless/b43legacy/xmit.c
drivers/net/wireless/bcm43xx/Kconfig
drivers/net/wireless/bcm43xx/bcm43xx.h
drivers/net/wireless/bcm43xx/bcm43xx_main.c
drivers/net/wireless/bcm43xx/bcm43xx_pio.c
drivers/net/wireless/bcm43xx/bcm43xx_xmit.c
drivers/net/wireless/hostap/hostap_80211.h
drivers/net/wireless/hostap/hostap_80211_rx.c
drivers/net/wireless/hostap/hostap_ap.c
drivers/net/wireless/hostap/hostap_common.h
drivers/net/wireless/hostap/hostap_cs.c
drivers/net/wireless/hostap/hostap_download.c
drivers/net/wireless/hostap/hostap_hw.c
drivers/net/wireless/hostap/hostap_info.c
drivers/net/wireless/hostap/hostap_ioctl.c
drivers/net/wireless/hostap/hostap_main.c
drivers/net/wireless/hostap/hostap_pci.c
drivers/net/wireless/hostap/hostap_wlan.h
drivers/net/wireless/ipw2100.c
drivers/net/wireless/ipw2200.c
drivers/net/wireless/ipw2200.h
drivers/net/wireless/iwlwifi/Kconfig
drivers/net/wireless/iwlwifi/iwl-3945-commands.h [moved from drivers/net/wireless/iwlwifi/iwl-commands.h with 61% similarity]
drivers/net/wireless/iwlwifi/iwl-3945-debug.h [copied from drivers/net/wireless/iwlwifi/iwl-debug.h with 75% similarity]
drivers/net/wireless/iwlwifi/iwl-3945-hw.h
drivers/net/wireless/iwlwifi/iwl-3945-io.h [new file with mode: 0644]
drivers/net/wireless/iwlwifi/iwl-3945-rs.c
drivers/net/wireless/iwlwifi/iwl-3945-rs.h
drivers/net/wireless/iwlwifi/iwl-3945.c
drivers/net/wireless/iwlwifi/iwl-3945.h
drivers/net/wireless/iwlwifi/iwl-4965-commands.h [new file with mode: 0644]
drivers/net/wireless/iwlwifi/iwl-4965-debug.h [moved from drivers/net/wireless/iwlwifi/iwl-debug.h with 75% similarity]
drivers/net/wireless/iwlwifi/iwl-4965-hw.h
drivers/net/wireless/iwlwifi/iwl-4965-io.h [new file with mode: 0644]
drivers/net/wireless/iwlwifi/iwl-4965-rs.c
drivers/net/wireless/iwlwifi/iwl-4965-rs.h
drivers/net/wireless/iwlwifi/iwl-4965.c
drivers/net/wireless/iwlwifi/iwl-4965.h
drivers/net/wireless/iwlwifi/iwl-channel.h [deleted file]
drivers/net/wireless/iwlwifi/iwl-eeprom.h [deleted file]
drivers/net/wireless/iwlwifi/iwl-helpers.h
drivers/net/wireless/iwlwifi/iwl-hw.h [deleted file]
drivers/net/wireless/iwlwifi/iwl-io.h [deleted file]
drivers/net/wireless/iwlwifi/iwl-priv.h [deleted file]
drivers/net/wireless/iwlwifi/iwl-prph.h
drivers/net/wireless/iwlwifi/iwl3945-base.c
drivers/net/wireless/iwlwifi/iwl4965-base.c
drivers/net/wireless/iwlwifi/iwlwifi.h [deleted file]
drivers/net/wireless/libertas/11d.c
drivers/net/wireless/libertas/11d.h
drivers/net/wireless/libertas/README
drivers/net/wireless/libertas/assoc.c
drivers/net/wireless/libertas/assoc.h
drivers/net/wireless/libertas/cmd.c
drivers/net/wireless/libertas/cmd.h [new file with mode: 0644]
drivers/net/wireless/libertas/cmdresp.c
drivers/net/wireless/libertas/debugfs.c
drivers/net/wireless/libertas/debugfs.h
drivers/net/wireless/libertas/decl.h
drivers/net/wireless/libertas/defs.h
drivers/net/wireless/libertas/dev.h
drivers/net/wireless/libertas/ethtool.c
drivers/net/wireless/libertas/host.h
drivers/net/wireless/libertas/hostcmd.h
drivers/net/wireless/libertas/if_cs.c
drivers/net/wireless/libertas/if_sdio.c
drivers/net/wireless/libertas/if_sdio.h
drivers/net/wireless/libertas/if_usb.c
drivers/net/wireless/libertas/if_usb.h
drivers/net/wireless/libertas/join.c
drivers/net/wireless/libertas/join.h
drivers/net/wireless/libertas/main.c
drivers/net/wireless/libertas/rx.c
drivers/net/wireless/libertas/scan.c
drivers/net/wireless/libertas/scan.h
drivers/net/wireless/libertas/tx.c
drivers/net/wireless/libertas/types.h
drivers/net/wireless/libertas/wext.c
drivers/net/wireless/libertas/wext.h
drivers/net/wireless/orinoco.c
drivers/net/wireless/orinoco.h
drivers/net/wireless/p54common.c
drivers/net/wireless/p54pci.c
drivers/net/wireless/p54pci.h
drivers/net/wireless/prism54/isl_38xx.h
drivers/net/wireless/prism54/isl_ioctl.c
drivers/net/wireless/prism54/islpci_dev.c
drivers/net/wireless/prism54/islpci_dev.h
drivers/net/wireless/prism54/islpci_eth.c
drivers/net/wireless/prism54/islpci_eth.h
drivers/net/wireless/prism54/islpci_mgt.c
drivers/net/wireless/prism54/islpci_mgt.h
drivers/net/wireless/ray_cs.c
drivers/net/wireless/rt2x00/rt2400pci.c
drivers/net/wireless/rt2x00/rt2400pci.h
drivers/net/wireless/rt2x00/rt2500pci.c
drivers/net/wireless/rt2x00/rt2500pci.h
drivers/net/wireless/rt2x00/rt2500usb.c
drivers/net/wireless/rt2x00/rt2500usb.h
drivers/net/wireless/rt2x00/rt2x00.h
drivers/net/wireless/rt2x00/rt2x00config.c
drivers/net/wireless/rt2x00/rt2x00debug.c
drivers/net/wireless/rt2x00/rt2x00debug.h
drivers/net/wireless/rt2x00/rt2x00dev.c
drivers/net/wireless/rt2x00/rt2x00dump.h [new file with mode: 0644]
drivers/net/wireless/rt2x00/rt2x00firmware.c
drivers/net/wireless/rt2x00/rt2x00lib.h
drivers/net/wireless/rt2x00/rt2x00mac.c
drivers/net/wireless/rt2x00/rt2x00pci.c
drivers/net/wireless/rt2x00/rt2x00pci.h
drivers/net/wireless/rt2x00/rt2x00rfkill.c
drivers/net/wireless/rt2x00/rt2x00ring.h
drivers/net/wireless/rt2x00/rt2x00usb.c
drivers/net/wireless/rt2x00/rt2x00usb.h
drivers/net/wireless/rt2x00/rt61pci.c
drivers/net/wireless/rt2x00/rt61pci.h
drivers/net/wireless/rt2x00/rt73usb.c
drivers/net/wireless/rt2x00/rt73usb.h
drivers/net/wireless/rtl8180.h [new file with mode: 0644]
drivers/net/wireless/rtl8180_dev.c [new file with mode: 0644]
drivers/net/wireless/rtl8180_grf5101.c [new file with mode: 0644]
drivers/net/wireless/rtl8180_grf5101.h [new file with mode: 0644]
drivers/net/wireless/rtl8180_max2820.c [new file with mode: 0644]
drivers/net/wireless/rtl8180_max2820.h [new file with mode: 0644]
drivers/net/wireless/rtl8180_rtl8225.c [new file with mode: 0644]
drivers/net/wireless/rtl8180_rtl8225.h [new file with mode: 0644]
drivers/net/wireless/rtl8180_sa2400.c [new file with mode: 0644]
drivers/net/wireless/rtl8180_sa2400.h [new file with mode: 0644]
drivers/net/wireless/rtl8187.h
drivers/net/wireless/rtl8187_dev.c
drivers/net/wireless/rtl8187_rtl8225.c
drivers/net/wireless/rtl8187_rtl8225.h
drivers/net/wireless/rtl818x.h
drivers/net/wireless/wavelan.c
drivers/net/wireless/wavelan.p.h
drivers/net/wireless/wavelan_cs.c
drivers/net/wireless/wavelan_cs.p.h
drivers/net/wireless/zd1211rw/Kconfig
drivers/net/wireless/zd1211rw/Makefile
drivers/net/wireless/zd1211rw/zd_chip.c
drivers/net/wireless/zd1211rw/zd_chip.h
drivers/net/wireless/zd1211rw/zd_def.h
drivers/net/wireless/zd1211rw/zd_ieee80211.c
drivers/net/wireless/zd1211rw/zd_ieee80211.h
drivers/net/wireless/zd1211rw/zd_mac.c
drivers/net/wireless/zd1211rw/zd_mac.h
drivers/net/wireless/zd1211rw/zd_netdev.c [deleted file]
drivers/net/wireless/zd1211rw/zd_netdev.h [deleted file]
drivers/net/wireless/zd1211rw/zd_rf.c
drivers/net/wireless/zd1211rw/zd_rf.h
drivers/net/wireless/zd1211rw/zd_rf_al2230.c
drivers/net/wireless/zd1211rw/zd_rf_al7230b.c
drivers/net/wireless/zd1211rw/zd_rf_rf2959.c
drivers/net/wireless/zd1211rw/zd_rf_uw2453.c
drivers/net/wireless/zd1211rw/zd_usb.c
drivers/net/wireless/zd1211rw/zd_usb.h
drivers/net/xen-netfront.c
drivers/parisc/led.c
drivers/s390/net/qeth_main.c
drivers/scsi/qla4xxx/ql4_os.c
drivers/scsi/scsi_netlink.c
drivers/scsi/scsi_transport_iscsi.c
drivers/ssb/b43_pci_bridge.c
drivers/ssb/main.c
drivers/ssb/pci.c
drivers/ssb/pcmcia.c
fs/ecryptfs/netlink.c
fs/proc/proc_net.c
fs/splice.c
include/asm-powerpc/pasemi_dma.h [new file with mode: 0644]
include/linux/Kbuild
include/linux/atmbr2684.h
include/linux/atmdev.h
include/linux/can.h [new file with mode: 0644]
include/linux/can/Kbuild [new file with mode: 0644]
include/linux/can/bcm.h [new file with mode: 0644]
include/linux/can/core.h [new file with mode: 0644]
include/linux/can/error.h [new file with mode: 0644]
include/linux/can/raw.h [new file with mode: 0644]
include/linux/connector.h
include/linux/dccp.h
include/linux/ieee80211.h
include/linux/if.h
include/linux/if_addrlabel.h [new file with mode: 0644]
include/linux/if_arp.h
include/linux/if_ether.h
include/linux/if_frad.h
include/linux/if_shaper.h [deleted file]
include/linux/if_tr.h
include/linux/if_tun.h
include/linux/if_tunnel.h
include/linux/if_vlan.h
include/linux/in.h
include/linux/inetdevice.h
include/linux/net.h
include/linux/netfilter.h
include/linux/netfilter/Kbuild
include/linux/netfilter/nf_conntrack_common.h
include/linux/netfilter/nf_conntrack_h323.h
include/linux/netfilter/nf_conntrack_sctp.h
include/linux/netfilter/nfnetlink_conntrack.h
include/linux/netfilter/nfnetlink_log.h
include/linux/netfilter/x_tables.h
include/linux/netfilter/xt_CONNMARK.h
include/linux/netfilter/xt_DSCP.h
include/linux/netfilter/xt_MARK.h
include/linux/netfilter/xt_RATEEST.h [new file with mode: 0644]
include/linux/netfilter/xt_TCPOPTSTRIP.h [new file with mode: 0644]
include/linux/netfilter/xt_connlimit.h
include/linux/netfilter/xt_connmark.h
include/linux/netfilter/xt_conntrack.h
include/linux/netfilter/xt_dscp.h
include/linux/netfilter/xt_hashlimit.h
include/linux/netfilter/xt_iprange.h [new file with mode: 0644]
include/linux/netfilter/xt_mark.h
include/linux/netfilter/xt_owner.h [new file with mode: 0644]
include/linux/netfilter/xt_policy.h
include/linux/netfilter/xt_quota.h
include/linux/netfilter/xt_rateest.h [new file with mode: 0644]
include/linux/netfilter/xt_statistic.h
include/linux/netfilter/xt_string.h
include/linux/netfilter_arp/arp_tables.h
include/linux/netfilter_ipv4.h
include/linux/netfilter_ipv4/ip_tables.h
include/linux/netfilter_ipv4/ipt_CLUSTERIP.h
include/linux/netfilter_ipv4/ipt_addrtype.h
include/linux/netfilter_ipv4/ipt_iprange.h
include/linux/netfilter_ipv6.h
include/linux/netfilter_ipv6/ip6_tables.h
include/linux/netlink.h
include/linux/netpoll.h
include/linux/nl80211.h
include/linux/pci_ids.h
include/linux/pcounter.h [new file with mode: 0644]
include/linux/pkt_sched.h
include/linux/proc_fs.h
include/linux/rtnetlink.h
include/linux/seq_file.h
include/linux/skbuff.h
include/linux/snmp.h
include/linux/socket.h
include/linux/splice.h
include/linux/ssb/ssb.h
include/linux/ssb/ssb_regs.h
include/linux/sysctl.h
include/linux/tcp.h
include/linux/tty.h
include/linux/wireless.h
include/linux/xfrm.h
include/net/act_api.h
include/net/addrconf.h
include/net/af_unix.h
include/net/arp.h
include/net/bluetooth/rfcomm.h
include/net/cfg80211.h
include/net/checksum.h
include/net/dsfield.h
include/net/dst.h
include/net/fib_rules.h
include/net/flow.h
include/net/gen_stats.h
include/net/ieee80211.h
include/net/inet_ecn.h
include/net/inet_frag.h
include/net/inet_hashtables.h
include/net/inet_timewait_sock.h
include/net/ip.h
include/net/ip6_fib.h
include/net/ip6_route.h
include/net/ip_fib.h
include/net/ip_vs.h
include/net/ipip.h
include/net/ipv6.h
include/net/irda/irda_device.h
include/net/mac80211.h
include/net/neighbour.h
include/net/net_namespace.h
include/net/netevent.h
include/net/netfilter/ipv6/nf_conntrack_ipv6.h
include/net/netfilter/nf_conntrack.h
include/net/netfilter/nf_conntrack_core.h
include/net/netfilter/nf_conntrack_expect.h
include/net/netfilter/nf_conntrack_helper.h
include/net/netfilter/nf_conntrack_l3proto.h
include/net/netfilter/nf_conntrack_tuple.h
include/net/netfilter/nf_log.h [new file with mode: 0644]
include/net/netfilter/nf_nat.h
include/net/netfilter/nf_nat_protocol.h
include/net/netfilter/nf_queue.h [new file with mode: 0644]
include/net/netfilter/xt_rateest.h [new file with mode: 0644]
include/net/netlink.h
include/net/netns/ipv4.h [new file with mode: 0644]
include/net/netns/ipv6.h [new file with mode: 0644]
include/net/netns/packet.h [new file with mode: 0644]
include/net/netns/unix.h [new file with mode: 0644]
include/net/pkt_cls.h
include/net/pkt_sched.h
include/net/protocol.h
include/net/raw.h
include/net/rawv6.h
include/net/route.h
include/net/sch_generic.h
include/net/sctp/checksum.h [new file with mode: 0644]
include/net/sctp/constants.h
include/net/sctp/sctp.h
include/net/sctp/structs.h
include/net/snmp.h
include/net/sock.h
include/net/tcp.h
include/net/transp_v6.h
include/net/udp.h
include/net/udplite.h
include/net/xfrm.h
kernel/sysctl.c
kernel/sysctl_check.c
lib/Makefile
lib/pcounter.c [new file with mode: 0644]
net/802/Makefile
net/802/sysctl_net_802.c [deleted file]
net/802/tr.c
net/8021q/vlan.c
net/8021q/vlan.h
net/8021q/vlan_dev.c
net/8021q/vlan_netlink.c
net/8021q/vlanproc.c
net/8021q/vlanproc.h
net/Kconfig
net/Makefile
net/appletalk/aarp.c
net/appletalk/atalk_proc.c
net/appletalk/ddp.c
net/appletalk/sysctl_net_atalk.c
net/atm/Kconfig
net/atm/atm_sysfs.c
net/atm/br2684.c
net/atm/clip.c
net/atm/common.c
net/atm/lec.c
net/atm/proc.c
net/ax25/af_ax25.c
net/ax25/ax25_ds_timer.c
net/ax25/ax25_route.c
net/ax25/ax25_std_timer.c
net/ax25/ax25_uid.c
net/ax25/sysctl_net_ax25.c
net/bluetooth/bnep/sock.c
net/bluetooth/cmtp/sock.c
net/bluetooth/hci_conn.c
net/bluetooth/hidp/core.c
net/bluetooth/hidp/sock.c
net/bluetooth/l2cap.c
net/bluetooth/rfcomm/core.c
net/bluetooth/sco.c
net/bridge/br_input.c
net/bridge/br_netfilter.c
net/bridge/br_netlink.c
net/bridge/netfilter/Kconfig
net/bridge/netfilter/ebt_log.c
net/bridge/netfilter/ebt_ulog.c
net/bridge/netfilter/ebt_vlan.c
net/bridge/netfilter/ebtable_filter.c
net/bridge/netfilter/ebtable_nat.c
net/bridge/netfilter/ebtables.c
net/can/Kconfig [new file with mode: 0644]
net/can/Makefile [new file with mode: 0644]
net/can/af_can.c [new file with mode: 0644]
net/can/af_can.h [new file with mode: 0644]
net/can/bcm.c [new file with mode: 0644]
net/can/proc.c [new file with mode: 0644]
net/can/raw.c [new file with mode: 0644]
net/compat.c
net/core/datagram.c
net/core/dev.c
net/core/dev_mcast.c
net/core/dst.c
net/core/fib_rules.c
net/core/flow.c
net/core/gen_estimator.c
net/core/gen_stats.c
net/core/neighbour.c
net/core/net-sysfs.c
net/core/net_namespace.c
net/core/netpoll.c
net/core/pktgen.c
net/core/request_sock.c
net/core/rtnetlink.c
net/core/skbuff.c
net/core/sock.c
net/core/stream.c
net/core/sysctl_net_core.c
net/core/utils.c
net/dccp/Kconfig
net/dccp/ackvec.c
net/dccp/ackvec.h
net/dccp/ccid.c
net/dccp/ccid.h
net/dccp/ccids/Kconfig
net/dccp/ccids/ccid2.c
net/dccp/ccids/ccid2.h
net/dccp/ccids/ccid3.c
net/dccp/ccids/ccid3.h
net/dccp/ccids/lib/Makefile
net/dccp/ccids/lib/loss_interval.c
net/dccp/ccids/lib/loss_interval.h
net/dccp/ccids/lib/packet_history.c
net/dccp/ccids/lib/packet_history.h
net/dccp/ccids/lib/tfrc.c [new file with mode: 0644]
net/dccp/ccids/lib/tfrc.h
net/dccp/dccp.h
net/dccp/feat.c
net/dccp/feat.h
net/dccp/input.c
net/dccp/ipv4.c
net/dccp/ipv6.c
net/dccp/minisocks.c
net/dccp/options.c
net/dccp/output.c
net/dccp/proto.c
net/dccp/sysctl.c
net/dccp/timer.c
net/decnet/af_decnet.c
net/decnet/dn_dev.c
net/decnet/dn_fib.c
net/decnet/dn_neigh.c
net/decnet/dn_nsp_out.c
net/decnet/dn_route.c
net/decnet/dn_rules.c
net/decnet/dn_table.c
net/decnet/netfilter/Kconfig
net/decnet/netfilter/dn_rtmsg.c
net/decnet/sysctl_net_decnet.c
net/econet/af_econet.c
net/ethernet/eth.c
net/ieee80211/Kconfig
net/ieee80211/ieee80211_crypt_tkip.c
net/ieee80211/ieee80211_module.c
net/ieee80211/ieee80211_rx.c
net/ieee80211/ieee80211_tx.c
net/ieee80211/ieee80211_wx.c
net/ieee80211/softmac/ieee80211softmac_auth.c
net/ieee80211/softmac/ieee80211softmac_io.c
net/ipv4/Kconfig
net/ipv4/Makefile
net/ipv4/af_inet.c
net/ipv4/ah4.c
net/ipv4/arp.c
net/ipv4/cipso_ipv4.c
net/ipv4/datagram.c
net/ipv4/devinet.c
net/ipv4/esp4.c
net/ipv4/fib_frontend.c
net/ipv4/fib_hash.c
net/ipv4/fib_lookup.h
net/ipv4/fib_rules.c
net/ipv4/fib_semantics.c
net/ipv4/fib_trie.c
net/ipv4/icmp.c
net/ipv4/igmp.c
net/ipv4/inet_connection_sock.c
net/ipv4/inet_diag.c
net/ipv4/inet_fragment.c
net/ipv4/inet_hashtables.c
net/ipv4/inet_timewait_sock.c
net/ipv4/ip_forward.c
net/ipv4/ip_fragment.c
net/ipv4/ip_gre.c
net/ipv4/ip_input.c
net/ipv4/ip_options.c
net/ipv4/ip_output.c
net/ipv4/ip_sockglue.c
net/ipv4/ipcomp.c
net/ipv4/ipconfig.c
net/ipv4/ipip.c
net/ipv4/ipmr.c
net/ipv4/ipvs/ip_vs_app.c
net/ipv4/ipvs/ip_vs_conn.c
net/ipv4/ipvs/ip_vs_core.c
net/ipv4/ipvs/ip_vs_ctl.c
net/ipv4/ipvs/ip_vs_est.c
net/ipv4/ipvs/ip_vs_lblc.c
net/ipv4/ipvs/ip_vs_lblcr.c
net/ipv4/ipvs/ip_vs_proto.c
net/ipv4/ipvs/ip_vs_proto_esp.c
net/ipv4/ipvs/ip_vs_sched.c
net/ipv4/ipvs/ip_vs_sync.c
net/ipv4/ipvs/ip_vs_xmit.c
net/ipv4/netfilter.c
net/ipv4/netfilter/Kconfig
net/ipv4/netfilter/Makefile
net/ipv4/netfilter/arp_tables.c
net/ipv4/netfilter/arptable_filter.c
net/ipv4/netfilter/ip_queue.c
net/ipv4/netfilter/ip_tables.c
net/ipv4/netfilter/ipt_CLUSTERIP.c
net/ipv4/netfilter/ipt_ECN.c
net/ipv4/netfilter/ipt_LOG.c
net/ipv4/netfilter/ipt_MASQUERADE.c
net/ipv4/netfilter/ipt_NETMAP.c
net/ipv4/netfilter/ipt_REDIRECT.c
net/ipv4/netfilter/ipt_REJECT.c
net/ipv4/netfilter/ipt_SAME.c [deleted file]
net/ipv4/netfilter/ipt_TOS.c [deleted file]
net/ipv4/netfilter/ipt_TTL.c
net/ipv4/netfilter/ipt_ULOG.c
net/ipv4/netfilter/ipt_addrtype.c
net/ipv4/netfilter/ipt_ah.c
net/ipv4/netfilter/ipt_ecn.c
net/ipv4/netfilter/ipt_iprange.c [deleted file]
net/ipv4/netfilter/ipt_owner.c [deleted file]
net/ipv4/netfilter/ipt_recent.c
net/ipv4/netfilter/ipt_tos.c [deleted file]
net/ipv4/netfilter/ipt_ttl.c
net/ipv4/netfilter/iptable_filter.c
net/ipv4/netfilter/iptable_mangle.c
net/ipv4/netfilter/iptable_raw.c
net/ipv4/netfilter/nf_conntrack_l3proto_ipv4.c
net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
net/ipv4/netfilter/nf_conntrack_proto_icmp.c
net/ipv4/netfilter/nf_nat_core.c
net/ipv4/netfilter/nf_nat_h323.c
net/ipv4/netfilter/nf_nat_helper.c
net/ipv4/netfilter/nf_nat_pptp.c
net/ipv4/netfilter/nf_nat_proto_gre.c
net/ipv4/netfilter/nf_nat_proto_icmp.c
net/ipv4/netfilter/nf_nat_proto_tcp.c
net/ipv4/netfilter/nf_nat_proto_udp.c
net/ipv4/netfilter/nf_nat_proto_unknown.c
net/ipv4/netfilter/nf_nat_rule.c
net/ipv4/netfilter/nf_nat_sip.c
net/ipv4/netfilter/nf_nat_snmp_basic.c
net/ipv4/netfilter/nf_nat_standalone.c
net/ipv4/proc.c
net/ipv4/raw.c
net/ipv4/route.c
net/ipv4/syncookies.c
net/ipv4/sysctl_net_ipv4.c
net/ipv4/tcp.c
net/ipv4/tcp_bic.c
net/ipv4/tcp_cong.c
net/ipv4/tcp_cubic.c
net/ipv4/tcp_highspeed.c
net/ipv4/tcp_htcp.c
net/ipv4/tcp_hybla.c
net/ipv4/tcp_illinois.c
net/ipv4/tcp_input.c
net/ipv4/tcp_ipv4.c
net/ipv4/tcp_lp.c
net/ipv4/tcp_output.c
net/ipv4/tcp_scalable.c
net/ipv4/tcp_timer.c
net/ipv4/tcp_vegas.c
net/ipv4/tcp_veno.c
net/ipv4/tcp_yeah.c
net/ipv4/udp.c
net/ipv4/udplite.c
net/ipv4/xfrm4_input.c
net/ipv4/xfrm4_mode_beet.c
net/ipv4/xfrm4_mode_tunnel.c
net/ipv4/xfrm4_output.c
net/ipv4/xfrm4_policy.c
net/ipv4/xfrm4_state.c
net/ipv6/Makefile
net/ipv6/addrconf.c
net/ipv6/addrlabel.c [new file with mode: 0644]
net/ipv6/af_inet6.c
net/ipv6/ah6.c
net/ipv6/anycast.c
net/ipv6/datagram.c
net/ipv6/esp6.c
net/ipv6/exthdrs.c
net/ipv6/fib6_rules.c
net/ipv6/icmp.c
net/ipv6/inet6_hashtables.c
net/ipv6/ip6_fib.c
net/ipv6/ip6_flowlabel.c
net/ipv6/ip6_input.c
net/ipv6/ip6_output.c
net/ipv6/ip6_tunnel.c
net/ipv6/ipcomp6.c
net/ipv6/ipv6_sockglue.c
net/ipv6/mcast.c
net/ipv6/mip6.c
net/ipv6/ndisc.c
net/ipv6/netfilter.c
net/ipv6/netfilter/Kconfig
net/ipv6/netfilter/Makefile
net/ipv6/netfilter/ip6_queue.c
net/ipv6/netfilter/ip6_tables.c
net/ipv6/netfilter/ip6t_HL.c
net/ipv6/netfilter/ip6t_LOG.c
net/ipv6/netfilter/ip6t_REJECT.c
net/ipv6/netfilter/ip6t_ah.c
net/ipv6/netfilter/ip6t_eui64.c
net/ipv6/netfilter/ip6t_frag.c
net/ipv6/netfilter/ip6t_hbh.c
net/ipv6/netfilter/ip6t_hl.c
net/ipv6/netfilter/ip6t_ipv6header.c
net/ipv6/netfilter/ip6t_mh.c
net/ipv6/netfilter/ip6t_owner.c [deleted file]
net/ipv6/netfilter/ip6t_rt.c
net/ipv6/netfilter/ip6table_filter.c
net/ipv6/netfilter/ip6table_mangle.c
net/ipv6/netfilter/ip6table_raw.c
net/ipv6/netfilter/nf_conntrack_l3proto_ipv6.c
net/ipv6/netfilter/nf_conntrack_proto_icmpv6.c
net/ipv6/netfilter/nf_conntrack_reasm.c
net/ipv6/proc.c
net/ipv6/raw.c
net/ipv6/reassembly.c
net/ipv6/route.c
net/ipv6/sit.c
net/ipv6/sysctl_net_ipv6.c
net/ipv6/tcp_ipv6.c
net/ipv6/udp.c
net/ipv6/udp_impl.h
net/ipv6/udplite.c
net/ipv6/xfrm6_input.c
net/ipv6/xfrm6_mode_beet.c
net/ipv6/xfrm6_mode_ro.c
net/ipv6/xfrm6_mode_tunnel.c
net/ipv6/xfrm6_output.c
net/ipv6/xfrm6_policy.c
net/ipv6/xfrm6_state.c
net/ipx/sysctl_net_ipx.c
net/irda/af_irda.c
net/irda/ircomm/ircomm_core.c
net/irda/irda_device.c
net/irda/iriap.c
net/irda/irlap_event.c
net/irda/irlmp.c
net/irda/irlmp_event.c
net/irda/irsysctl.c
net/iucv/af_iucv.c
net/iucv/iucv.c
net/key/af_key.c
net/lapb/lapb_iface.c
net/llc/llc_conn.c
net/llc/llc_station.c
net/llc/sysctl_net_llc.c
net/mac80211/Kconfig
net/mac80211/Makefile
net/mac80211/cfg.c
net/mac80211/debugfs_netdev.c
net/mac80211/ieee80211.c
net/mac80211/ieee80211_i.h
net/mac80211/ieee80211_iface.c
net/mac80211/ieee80211_ioctl.c
net/mac80211/ieee80211_led.c
net/mac80211/ieee80211_led.h
net/mac80211/ieee80211_rate.c
net/mac80211/ieee80211_rate.h
net/mac80211/ieee80211_sta.c
net/mac80211/key.c
net/mac80211/rc80211_pid.h [new file with mode: 0644]
net/mac80211/rc80211_pid_algo.c [new file with mode: 0644]
net/mac80211/rc80211_pid_debugfs.c [new file with mode: 0644]
net/mac80211/rc80211_simple.c
net/mac80211/rx.c
net/mac80211/sta_info.c
net/mac80211/sta_info.h
net/mac80211/tx.c
net/mac80211/util.c
net/mac80211/wep.c
net/mac80211/wme.c
net/mac80211/wpa.c
net/netfilter/Kconfig
net/netfilter/Makefile
net/netfilter/core.c
net/netfilter/nf_conntrack_core.c
net/netfilter/nf_conntrack_expect.c
net/netfilter/nf_conntrack_ftp.c
net/netfilter/nf_conntrack_h323_asn1.c
net/netfilter/nf_conntrack_h323_main.c
net/netfilter/nf_conntrack_l3proto_generic.c
net/netfilter/nf_conntrack_netlink.c
net/netfilter/nf_conntrack_proto.c
net/netfilter/nf_conntrack_proto_generic.c
net/netfilter/nf_conntrack_proto_sctp.c
net/netfilter/nf_conntrack_proto_tcp.c
net/netfilter/nf_conntrack_proto_udp.c
net/netfilter/nf_conntrack_proto_udplite.c
net/netfilter/nf_conntrack_sip.c
net/netfilter/nf_conntrack_standalone.c
net/netfilter/nf_log.c
net/netfilter/nf_queue.c
net/netfilter/nf_sysctl.c [deleted file]
net/netfilter/nfnetlink.c
net/netfilter/nfnetlink_log.c
net/netfilter/nfnetlink_queue.c
net/netfilter/x_tables.c
net/netfilter/xt_CLASSIFY.c
net/netfilter/xt_CONNMARK.c
net/netfilter/xt_CONNSECMARK.c
net/netfilter/xt_DSCP.c
net/netfilter/xt_MARK.c
net/netfilter/xt_NFLOG.c
net/netfilter/xt_NFQUEUE.c
net/netfilter/xt_NOTRACK.c
net/netfilter/xt_RATEEST.c [new file with mode: 0644]
net/netfilter/xt_SECMARK.c
net/netfilter/xt_TCPMSS.c
net/netfilter/xt_TCPOPTSTRIP.c [new file with mode: 0644]
net/netfilter/xt_TRACE.c
net/netfilter/xt_comment.c
net/netfilter/xt_connbytes.c
net/netfilter/xt_connlimit.c
net/netfilter/xt_connmark.c
net/netfilter/xt_conntrack.c
net/netfilter/xt_dccp.c
net/netfilter/xt_dscp.c
net/netfilter/xt_esp.c
net/netfilter/xt_hashlimit.c
net/netfilter/xt_helper.c
net/netfilter/xt_iprange.c [new file with mode: 0644]
net/netfilter/xt_length.c
net/netfilter/xt_limit.c
net/netfilter/xt_mac.c
net/netfilter/xt_mark.c
net/netfilter/xt_multiport.c
net/netfilter/xt_owner.c [new file with mode: 0644]
net/netfilter/xt_physdev.c
net/netfilter/xt_pkttype.c
net/netfilter/xt_policy.c
net/netfilter/xt_quota.c
net/netfilter/xt_rateest.c [new file with mode: 0644]
net/netfilter/xt_realm.c
net/netfilter/xt_sctp.c
net/netfilter/xt_state.c
net/netfilter/xt_statistic.c
net/netfilter/xt_string.c
net/netfilter/xt_tcpmss.c
net/netfilter/xt_tcpudp.c
net/netfilter/xt_time.c
net/netfilter/xt_u32.c
net/netlink/af_netlink.c
net/netlink/attr.c
net/netrom/nr_timer.c
net/netrom/sysctl_net_netrom.c
net/packet/af_packet.c
net/rose/af_rose.c
net/rose/rose_in.c
net/rose/rose_route.c
net/rose/sysctl_net_rose.c
net/rxrpc/af_rxrpc.c
net/rxrpc/ar-connection.c
net/rxrpc/ar-input.c
net/rxrpc/ar-peer.c
net/rxrpc/rxkad.c
net/sched/Kconfig
net/sched/act_api.c
net/sched/act_gact.c
net/sched/act_ipt.c
net/sched/act_mirred.c
net/sched/act_nat.c
net/sched/act_pedit.c
net/sched/act_police.c
net/sched/act_simple.c
net/sched/cls_api.c
net/sched/cls_basic.c
net/sched/cls_fw.c
net/sched/cls_route.c
net/sched/cls_rsvp.h
net/sched/cls_tcindex.c
net/sched/cls_u32.c
net/sched/em_meta.c
net/sched/em_text.c
net/sched/ematch.c
net/sched/sch_api.c
net/sched/sch_atm.c
net/sched/sch_blackhole.c
net/sched/sch_cbq.c
net/sched/sch_dsmark.c
net/sched/sch_fifo.c
net/sched/sch_generic.c
net/sched/sch_gred.c
net/sched/sch_hfsc.c
net/sched/sch_htb.c
net/sched/sch_ingress.c
net/sched/sch_netem.c
net/sched/sch_prio.c
net/sched/sch_red.c
net/sched/sch_sfq.c
net/sched/sch_tbf.c
net/sched/sch_teql.c
net/sctp/Kconfig
net/sctp/Makefile
net/sctp/associola.c
net/sctp/bind_addr.c
net/sctp/crc32c.c [deleted file]
net/sctp/input.c
net/sctp/ipv6.c
net/sctp/output.c
net/sctp/outqueue.c
net/sctp/protocol.c
net/sctp/sm_make_chunk.c
net/sctp/sm_statefuns.c
net/sctp/sm_statetable.c
net/sctp/socket.c
net/sctp/sysctl.c
net/sctp/transport.c
net/sctp/ulpevent.c
net/sctp/ulpqueue.c
net/socket.c
net/sunrpc/cache.c
net/sunrpc/sched.c
net/sunrpc/xprt.c
net/sunrpc/xprtrdma/rpc_rdma.c
net/sunrpc/xprtsock.c
net/sysctl_net.c
net/tipc/core.h
net/tipc/port.c
net/unix/af_unix.c
net/unix/sysctl_net_unix.c
net/wireless/Kconfig
net/wireless/core.c
net/wireless/nl80211.c
net/wireless/wext.c
net/x25/af_x25.c
net/x25/sysctl_net_x25.c
net/x25/x25_facilities.c
net/x25/x25_forward.c
net/x25/x25_in.c
net/x25/x25_link.c
net/x25/x25_proc.c
net/x25/x25_route.c
net/x25/x25_subr.c
net/x25/x25_timer.c
net/xfrm/Kconfig
net/xfrm/Makefile
net/xfrm/xfrm_algo.c
net/xfrm/xfrm_hash.c
net/xfrm/xfrm_input.c
net/xfrm/xfrm_output.c
net/xfrm/xfrm_policy.c
net/xfrm/xfrm_proc.c [new file with mode: 0644]
net/xfrm/xfrm_state.c
net/xfrm/xfrm_user.c
security/selinux/hooks.c

diff --git a/CREDITS b/CREDITS
index ee909f2..edff310 100644 (file)
--- a/CREDITS
+++ b/CREDITS
@@ -1353,6 +1353,14 @@ S: Gen Stedmanstraat 212
 S: 5623 HZ Eindhoven
 S: The Netherlands
 
+N: Oliver Hartkopp
+E: oliver.hartkopp@volkswagen.de
+W: http://www.volkswagen.de
+D: Controller Area Network (network layer core)
+S: Brieffach 1776
+S: 38436 Wolfsburg
+S: Germany
+
 N: Andrew Haylett
 E: ajh@primag.co.uk
 D: Selection mechanism
@@ -3306,6 +3314,14 @@ S: Universit=E9 de Rennes I
 S: F-35042 Rennes Cedex
 S: France
 
+N: Urs Thuermann
+E: urs.thuermann@volkswagen.de
+W: http://www.volkswagen.de
+D: Controller Area Network (network layer core)
+S: Brieffach 1776
+S: 38436 Wolfsburg
+S: Germany
+
 N: Jon Tombs
 E: jon@gte.esi.us.es
 W: http://www.esi.us.es/~jon
index 2537066..181bff0 100644 (file)
@@ -249,15 +249,6 @@ Who:  Tejun Heo <htejun@gmail.com>
 
 ---------------------------
 
-What:  iptables SAME target
-When:  1.1. 2008
-Files: net/ipv4/netfilter/ipt_SAME.c, include/linux/netfilter_ipv4/ipt_SAME.h
-Why:   Obsolete for multiple years now, NAT core provides the same behaviour.
-       Unfixable broken wrt. 32/64 bit cleanness.
-Who:   Patrick McHardy <kaber@trash.net>
-
----------------------------
-
 What: The arch/ppc and include/asm-ppc directories
 When: Jun 2008
 Why:  The arch/powerpc tree is the merged architecture for ppc32 and ppc64
@@ -289,15 +280,6 @@ Who:       Thomas Gleixner <tglx@linutronix.de>
 
 ---------------------------
 
-What:  shaper network driver
-When:  January 2008
-Files: drivers/net/shaper.c, include/linux/if_shaper.h
-Why:   This driver has been marked obsolete for many years.
-       It was only designed to work on lower speed links and has design
-       flaws that lead to machine crashes. The qdisc infrastructure in
-       2.4 or later kernels, provides richer features and is more robust.
-Who:   Stephen Hemminger <shemminger@linux-foundation.org>
-
 ---------------------------
 
 What:  i2c-i810, i2c-prosavage and i2c-savage4
@@ -306,3 +288,69 @@ Why:       These drivers are superseded by i810fb, intelfb and savagefb.
 Who:   Jean Delvare <khali@linux-fr.org>
 
 ---------------------------
+
+What:  bcm43xx wireless network driver
+When:  2.6.26
+Files: drivers/net/wireless/bcm43xx
+Why:   This driver's functionality has been replaced by the
+       mac80211-based b43 and b43legacy drivers.
+Who:   John W. Linville <linville@tuxdriver.com>
+
+---------------------------
+
+What:  ieee80211 softmac wireless networking component
+When:  2.6.26 (or after removal of bcm43xx and port of zd1211rw to mac80211)
+Files: net/ieee80211/softmac
+Why:   No in-kernel drivers will depend on it any longer.
+Who:   John W. Linville <linville@tuxdriver.com>
+
+---------------------------
+
+What:  rc80211-simple rate control algorithm for mac80211
+When:  2.6.26
+Files: net/mac80211/rc80211-simple.c
+Why:   This algorithm was provided for reference but always exhibited bad
+       responsiveness and performance and has some serious flaws. It has been
+       replaced by rc80211-pid.
+Who:   Stefano Brivio <stefano.brivio@polimi.it>
+
+---------------------------
+
+What (Why):
+       - include/linux/netfilter_ipv4/ipt_TOS.h ipt_tos.h header files
+         (superseded by xt_TOS/xt_tos target & match)
+
+       - "forwarding" header files like ipt_mac.h in
+         include/linux/netfilter_ipv4/ and include/linux/netfilter_ipv6/
+
+       - xt_CONNMARK match revision 0
+         (superseded by xt_CONNMARK match revision 1)
+
+       - xt_MARK target revisions 0 and 1
+         (superseded by xt_MARK match revision 2)
+
+       - xt_connmark match revision 0
+         (superseded by xt_connmark match revision 1)
+
+       - xt_conntrack match revision 0
+         (superseded by xt_conntrack match revision 1)
+
+       - xt_iprange match revision 0,
+         include/linux/netfilter_ipv4/ipt_iprange.h
+         (superseded by xt_iprange match revision 1)
+
+       - xt_mark match revision 0
+         (superseded by xt_mark match revision 1)
+
+When:  January 2009 or Linux 2.7.0, whichever comes first
+Why:   Superseded by newer revisions or modules
+Who:   Jan Engelhardt <jengelh@computergmbh.de>
+
+---------------------------
+
+What:  b43 support for firmware revision < 410
+When:  July 2008
+Why:   The support code for the old firmware hurts code readability/maintainability
+       and slightly hurts runtime performance. Bugfixes for the old firmware
+       are not provided by Broadcom anymore.
+Who:   Michael Buesch <mb@bu3sch.de>
index 563e442..02e56d4 100644 (file)
@@ -24,6 +24,8 @@ baycom.txt
        - info on the driver for Baycom style amateur radio modems
 bridge.txt
        - where to get user space programs for ethernet bridging with Linux.
+can.txt
+       - documentation on CAN protocol family.
 cops.txt
        - info on the COPS LocalTalk Linux driver
 cs89x0.txt
@@ -82,8 +84,6 @@ policy-routing.txt
        - IP policy-based routing
 ray_cs.txt
        - Raylink Wireless LAN card driver info.
-shaper.txt
-       - info on the module that can shape/limit transmitted traffic.
 sk98lin.txt
        - Marvell Yukon Chipset / SysKonnect SK-98xx compliant Gigabit
          Ethernet Adapter family driver info
index 6cc30e0..a0cda06 100644 (file)
@@ -1,7 +1,7 @@
 
                Linux Ethernet Bonding Driver HOWTO
 
-               Latest update: 24 April 2006
+               Latest update: 12 November 2007
 
 Initial release : Thomas Davis <tadavis at lbl.gov>
 Corrections, HA extensions : 2000/10/03-15 :
@@ -166,12 +166,17 @@ to use ifenslave.
 2. Bonding Driver Options
 =========================
 
-       Options for the bonding driver are supplied as parameters to
-the bonding module at load time.  They may be given as command line
-arguments to the insmod or modprobe command, but are usually specified
-in either the /etc/modules.conf or /etc/modprobe.conf configuration
-file, or in a distro-specific configuration file (some of which are
-detailed in the next section).
+       Options for the bonding driver are supplied as parameters to the
+bonding module at load time, or are specified via sysfs.
+
+       Module options may be given as command line arguments to the
+insmod or modprobe command, but are usually specified in either the
+/etc/modules.conf or /etc/modprobe.conf configuration file, or in a
+distro-specific configuration file (some of which are detailed in the next
+section).
+
+       Details on bonding support for sysfs is provided in the
+"Configuring Bonding Manually via Sysfs" section, below.
 
        The available bonding driver parameters are listed below. If a
 parameter is not specified the default value is used.  When initially
@@ -812,11 +817,13 @@ the system /etc/modules.conf or /etc/modprobe.conf configuration file.
 3.2 Configuration with Initscripts Support
 ------------------------------------------
 
-       This section applies to distros using a version of initscripts
-with bonding support, for example, Red Hat Linux 9 or Red Hat
-Enterprise Linux version 3 or 4.  On these systems, the network
-initialization scripts have some knowledge of bonding, and can be
-configured to control bonding devices.
+       This section applies to distros using a recent version of
+initscripts with bonding support, for example, Red Hat Enterprise Linux
+version 3 or later, Fedora, etc.  On these systems, the network
+initialization scripts have knowledge of bonding, and can be configured to
+control bonding devices.  Note that older versions of the initscripts
+package have lower levels of support for bonding; this will be noted where
+applicable.
 
        These distros will not automatically load the network adapter
 driver unless the ethX device is configured with an IP address.
@@ -864,11 +871,31 @@ USERCTL=no
        Be sure to change the networking specific lines (IPADDR,
 NETMASK, NETWORK and BROADCAST) to match your network configuration.
 
-       Finally, it is necessary to edit /etc/modules.conf (or
-/etc/modprobe.conf, depending upon your distro) to load the bonding
-module with your desired options when the bond0 interface is brought
-up.  The following lines in /etc/modules.conf (or modprobe.conf) will
-load the bonding module, and select its options:
+       For later versions of initscripts, such as that found with Fedora
+7 and Red Hat Enterprise Linux version 5 (or later), it is possible, and,
+indeed, preferable, to specify the bonding options in the ifcfg-bond0
+file, e.g. a line of the format:
+
+BONDING_OPTS="mode=active-backup arp_interval=60 arp_ip_target=+192.168.1.254"
+
+       will configure the bond with the specified options.  The options
+specified in BONDING_OPTS are identical to the bonding module parameters
+except for the arp_ip_target field.  Each target should be included as a
+separate option and should be preceded by a '+' to indicate it should be
+added to the list of queried targets, e.g.,
+
+       arp_ip_target=+192.168.1.1 arp_ip_target=+192.168.1.2
+
+       is the proper syntax to specify multiple targets.  When specifying
+options via BONDING_OPTS, it is not necessary to edit /etc/modules.conf or
+/etc/modprobe.conf.
+
+       For older versions of initscripts that do not support
+BONDING_OPTS, it is necessary to edit /etc/modules.conf (or
+/etc/modprobe.conf, depending upon your distro) to load the bonding module
+with your desired options when the bond0 interface is brought up.  The
+following lines in /etc/modules.conf (or modprobe.conf) will load the
+bonding module, and select its options:
 
 alias bond0 bonding
 options bond0 mode=balance-alb miimon=100
@@ -883,9 +910,10 @@ up and running.
 3.2.1 Using DHCP with Initscripts
 ---------------------------------
 
-       Recent versions of initscripts (the version supplied with
-Fedora Core 3 and Red Hat Enterprise Linux 4 is reported to work) do
-have support for assigning IP information to bonding devices via DHCP.
+       Recent versions of initscripts (the versions supplied with Fedora
+Core 3 and Red Hat Enterprise Linux 4, or later versions, are reported to
+work) have support for assigning IP information to bonding devices via
+DHCP.
 
        To configure bonding for DHCP, configure it as described
 above, except replace the line "BOOTPROTO=none" with "BOOTPROTO=dhcp"
@@ -895,18 +923,14 @@ is case sensitive.
 3.2.2 Configuring Multiple Bonds with Initscripts
 -------------------------------------------------
 
-       At this writing, the initscripts package does not directly
-support loading the bonding driver multiple times, so the process for
-doing so is the same as described in the "Configuring Multiple Bonds
-Manually" section, below.
-
-       NOTE: It has been observed that some Red Hat supplied kernels
-are apparently unable to rename modules at load time (the "-o bond1"
-part).  Attempts to pass that option to modprobe will produce an
-"Operation not permitted" error.  This has been reported on some
-Fedora Core kernels, and has been seen on RHEL 4 as well.  On kernels
-exhibiting this problem, it will be impossible to configure multiple
-bonds with differing parameters.
+       Initscripts packages that are included with Fedora 7 and Red Hat
+Enterprise Linux 5 support multiple bonding interfaces by simply
+specifying the appropriate BONDING_OPTS= in ifcfg-bondX where X is the
+number of the bond.  This support requires sysfs support in the kernel,
+and a bonding driver of version 3.0.0 or later.  Other configurations may
+not support this method for specifying multiple bonding interfaces; for
+those instances, see the "Configuring Multiple Bonds Manually" section,
+below.
 
 3.3 Configuring Bonding Manually with Ifenslave
 -----------------------------------------------
@@ -977,15 +1001,58 @@ initialization scripts lack support for configuring multiple bonds.
 options, you may wish to use the "max_bonds" module parameter,
 documented above.
 
-       To create multiple bonding devices with differing options, it
-is necessary to use bonding parameters exported by sysfs, documented
-in the section below.
+       To create multiple bonding devices with differing options, it is
+preferrable to use bonding parameters exported by sysfs, documented in the
+section below.
+
+       For versions of bonding without sysfs support, the only means to
+provide multiple instances of bonding with differing options is to load
+the bonding driver multiple times.  Note that current versions of the
+sysconfig network initialization scripts handle this automatically; if
+your distro uses these scripts, no special action is needed.  See the
+section Configuring Bonding Devices, above, if you're not sure about your
+network initialization scripts.
+
+       To load multiple instances of the module, it is necessary to
+specify a different name for each instance (the module loading system
+requires that every loaded module, even multiple instances of the same
+module, have a unique name).  This is accomplished by supplying multiple
+sets of bonding options in /etc/modprobe.conf, for example:
+
+alias bond0 bonding
+options bond0 -o bond0 mode=balance-rr miimon=100
+
+alias bond1 bonding
+options bond1 -o bond1 mode=balance-alb miimon=50
+
+       will load the bonding module two times.  The first instance is
+named "bond0" and creates the bond0 device in balance-rr mode with an
+miimon of 100.  The second instance is named "bond1" and creates the
+bond1 device in balance-alb mode with an miimon of 50.
+
+       In some circumstances (typically with older distributions),
+the above does not work, and the second bonding instance never sees
+its options.  In that case, the second options line can be substituted
+as follows:
+
+install bond1 /sbin/modprobe --ignore-install bonding -o bond1 \
+       mode=balance-alb miimon=50
 
+       This may be repeated any number of times, specifying a new and
+unique name in place of bond1 for each subsequent instance.
+
+       It has been observed that some Red Hat supplied kernels are unable
+to rename modules at load time (the "-o bond1" part).  Attempts to pass
+that option to modprobe will produce an "Operation not permitted" error.
+This has been reported on some Fedora Core kernels, and has been seen on
+RHEL 4 as well.  On kernels exhibiting this problem, it will be impossible
+to configure multiple bonds with differing parameters (as they are older
+kernels, and also lack sysfs support).
 
 3.4 Configuring Bonding Manually via Sysfs
 ------------------------------------------
 
-       Starting with version 3.0, Channel Bonding may be configured
+       Starting with version 3.0.0, Channel Bonding may be configured
 via the sysfs interface.  This interface allows dynamic configuration
 of all bonds in the system without unloading the module.  It also
 allows for adding and removing bonds at runtime.  Ifenslave is no
@@ -1030,9 +1097,6 @@ To enslave interface eth0 to bond bond0:
 To free slave eth0 from bond bond0:
 # echo -eth0 > /sys/class/net/bond0/bonding/slaves
 
-       NOTE: The bond must be up before slaves can be added.  All
-slaves are freed when the interface is brought down.
-
        When an interface is enslaved to a bond, symlinks between the
 two are created in the sysfs filesystem.  In this case, you would get
 /sys/class/net/bond0/slave_eth0 pointing to /sys/class/net/eth0, and
@@ -1622,6 +1686,15 @@ one for each switch in the network).  This will insure that,
 regardless of which switch is active, the ARP monitor has a suitable
 target to query.
 
+       Note, also, that of late many switches now support a functionality
+generally referred to as "trunk failover."  This is a feature of the
+switch that causes the link state of a particular switch port to be set
+down (or up) when the state of another switch port goes down (or up).
+It's purpose is to propogate link failures from logically "exterior" ports
+to the logically "interior" ports that bonding is able to monitor via
+miimon.  Availability and configuration for trunk failover varies by
+switch, but this can be a viable alternative to the ARP monitor when using
+suitable switches.
 
 12. Configuring Bonding for Maximum Throughput
 ==============================================
@@ -1709,7 +1782,7 @@ balance-rr: This mode is the only mode that will permit a single
        interfaces. It is therefore the only mode that will allow a
        single TCP/IP stream to utilize more than one interface's
        worth of throughput.  This comes at a cost, however: the
-       striping often results in peer systems receiving packets out
+       striping generally results in peer systems receiving packets out
        of order, causing TCP/IP's congestion control system to kick
        in, often by retransmitting segments.
 
@@ -1721,22 +1794,20 @@ balance-rr: This mode is the only mode that will permit a single
        interface's worth of throughput, even after adjusting
        tcp_reordering.
 
-       Note that this out of order delivery occurs when both the
-       sending and receiving systems are utilizing a multiple
-       interface bond.  Consider a configuration in which a
-       balance-rr bond feeds into a single higher capacity network
-       channel (e.g., multiple 100Mb/sec ethernets feeding a single
-       gigabit ethernet via an etherchannel capable switch).  In this
-       configuration, traffic sent from the multiple 100Mb devices to
-       a destination connected to the gigabit device will not see
-       packets out of order.  However, traffic sent from the gigabit
-       device to the multiple 100Mb devices may or may not see
-       traffic out of order, depending upon the balance policy of the
-       switch.  Many switches do not support any modes that stripe
-       traffic (instead choosing a port based upon IP or MAC level
-       addresses); for those devices, traffic flowing from the
-       gigabit device to the many 100Mb devices will only utilize one
-       interface.
+       Note that the fraction of packets that will be delivered out of
+       order is highly variable, and is unlikely to be zero.  The level
+       of reordering depends upon a variety of factors, including the
+       networking interfaces, the switch, and the topology of the
+       configuration.  Speaking in general terms, higher speed network
+       cards produce more reordering (due to factors such as packet
+       coalescing), and a "many to many" topology will reorder at a
+       higher rate than a "many slow to one fast" configuration.
+
+       Many switches do not support any modes that stripe traffic
+       (instead choosing a port based upon IP or MAC level addresses);
+       for those devices, traffic for a particular connection flowing
+       through the switch to a balance-rr bond will not utilize greater
+       than one interface's worth of bandwidth.
 
        If you are utilizing protocols other than TCP/IP, UDP for
        example, and your application can tolerate out of order
@@ -1936,6 +2007,10 @@ Failover may be delayed via the downdelay bonding module option.
 13.2 Duplicated Incoming Packets
 --------------------------------
 
+       NOTE: Starting with version 3.0.2, the bonding driver has logic to
+suppress duplicate packets, which should largely eliminate this problem.
+The following description is kept for reference.
+
        It is not uncommon to observe a short burst of duplicated
 traffic when the bonding device is first used, or after it has been
 idle for some period of time.  This is most easily observed by issuing
@@ -2096,6 +2171,9 @@ The new driver was designed to be SMP safe from the start.
 EtherExpress PRO/100 and a 3com 3c905b, for example).  For most modes,
 devices need not be of the same speed.
 
+       Starting with version 3.2.1, bonding also supports Infiniband
+slaves in active-backup mode.
+
 3.  How many bonding devices can I have?
 
        There is no limit.
@@ -2154,11 +2232,15 @@ switches currently available support 802.3ad.
 
 8.  Where does a bonding device get its MAC address from?
 
-       If not explicitly configured (with ifconfig or ip link), the
-MAC address of the bonding device is taken from its first slave
-device.  This MAC address is then passed to all following slaves and
-remains persistent (even if the first slave is removed) until the
-bonding device is brought down or reconfigured.
+       When using slave devices that have fixed MAC addresses, or when
+the fail_over_mac option is enabled, the bonding device's MAC address is
+the MAC address of the active slave.
+
+       For other configurations, if not explicitly configured (with
+ifconfig or ip link), the MAC address of the bonding device is taken from
+its first slave device.  This MAC address is then passed to all following
+slaves and remains persistent (even if the first slave is removed) until
+the bonding device is brought down or reconfigured.
 
        If you wish to change the MAC address, you can set it with
 ifconfig or ip link:
diff --git a/Documentation/networking/can.txt b/Documentation/networking/can.txt
new file mode 100644 (file)
index 0000000..f1b2de1
--- /dev/null
@@ -0,0 +1,629 @@
+============================================================================
+
+can.txt
+
+Readme file for the Controller Area Network Protocol Family (aka Socket CAN)
+
+This file contains
+
+  1 Overview / What is Socket CAN
+
+  2 Motivation / Why using the socket API
+
+  3 Socket CAN concept
+    3.1 receive lists
+    3.2 local loopback of sent frames
+    3.3 network security issues (capabilities)
+    3.4 network problem notifications
+
+  4 How to use Socket CAN
+    4.1 RAW protocol sockets with can_filters (SOCK_RAW)
+      4.1.1 RAW socket option CAN_RAW_FILTER
+      4.1.2 RAW socket option CAN_RAW_ERR_FILTER
+      4.1.3 RAW socket option CAN_RAW_LOOPBACK
+      4.1.4 RAW socket option CAN_RAW_RECV_OWN_MSGS
+    4.2 Broadcast Manager protocol sockets (SOCK_DGRAM)
+    4.3 connected transport protocols (SOCK_SEQPACKET)
+    4.4 unconnected transport protocols (SOCK_DGRAM)
+
+  5 Socket CAN core module
+    5.1 can.ko module params
+    5.2 procfs content
+    5.3 writing own CAN protocol modules
+
+  6 CAN network drivers
+    6.1 general settings
+    6.2 local loopback of sent frames
+    6.3 CAN controller hardware filters
+    6.4 currently supported CAN hardware
+    6.5 todo
+
+  7 Credits
+
+============================================================================
+
+1. Overview / What is Socket CAN
+--------------------------------
+
+The socketcan package is an implementation of CAN protocols
+(Controller Area Network) for Linux.  CAN is a networking technology
+which has widespread use in automation, embedded devices, and
+automotive fields.  While there have been other CAN implementations
+for Linux based on character devices, Socket CAN uses the Berkeley
+socket API, the Linux network stack and implements the CAN device
+drivers as network interfaces.  The CAN socket API has been designed
+as similar as possible to the TCP/IP protocols to allow programmers,
+familiar with network programming, to easily learn how to use CAN
+sockets.
+
+2. Motivation / Why using the socket API
+----------------------------------------
+
+There have been CAN implementations for Linux before Socket CAN so the
+question arises, why we have started another project.  Most existing
+implementations come as a device driver for some CAN hardware, they
+are based on character devices and provide comparatively little
+functionality.  Usually, there is only a hardware-specific device
+driver which provides a character device interface to send and
+receive raw CAN frames, directly to/from the controller hardware.
+Queueing of frames and higher-level transport protocols like ISO-TP
+have to be implemented in user space applications.  Also, most
+character-device implementations support only one single process to
+open the device at a time, similar to a serial interface.  Exchanging
+the CAN controller requires employment of another device driver and
+often the need for adaption of large parts of the application to the
+new driver's API.
+
+Socket CAN was designed to overcome all of these limitations.  A new
+protocol family has been implemented which provides a socket interface
+to user space applications and which builds upon the Linux network
+layer, so to use all of the provided queueing functionality.  A device
+driver for CAN controller hardware registers itself with the Linux
+network layer as a network device, so that CAN frames from the
+controller can be passed up to the network layer and on to the CAN
+protocol family module and also vice-versa.  Also, the protocol family
+module provides an API for transport protocol modules to register, so
+that any number of transport protocols can be loaded or unloaded
+dynamically.  In fact, the can core module alone does not provide any
+protocol and cannot be used without loading at least one additional
+protocol module.  Multiple sockets can be opened at the same time,
+on different or the same protocol module and they can listen/send
+frames on different or the same CAN IDs.  Several sockets listening on
+the same interface for frames with the same CAN ID are all passed the
+same received matching CAN frames.  An application wishing to
+communicate using a specific transport protocol, e.g. ISO-TP, just
+selects that protocol when opening the socket, and then can read and
+write application data byte streams, without having to deal with
+CAN-IDs, frames, etc.
+
+Similar functionality visible from user-space could be provided by a
+character device, too, but this would lead to a technically inelegant
+solution for a couple of reasons:
+
+* Intricate usage.  Instead of passing a protocol argument to
+  socket(2) and using bind(2) to select a CAN interface and CAN ID, an
+  application would have to do all these operations using ioctl(2)s.
+
+* Code duplication.  A character device cannot make use of the Linux
+  network queueing code, so all that code would have to be duplicated
+  for CAN networking.
+
+* Abstraction.  In most existing character-device implementations, the
+  hardware-specific device driver for a CAN controller directly
+  provides the character device for the application to work with.
+  This is at least very unusual in Unix systems for both, char and
+  block devices.  For example you don't have a character device for a
+  certain UART of a serial interface, a certain sound chip in your
+  computer, a SCSI or IDE controller providing access to your hard
+  disk or tape streamer device.  Instead, you have abstraction layers
+  which provide a unified character or block device interface to the
+  application on the one hand, and a interface for hardware-specific
+  device drivers on the other hand.  These abstractions are provided
+  by subsystems like the tty layer, the audio subsystem or the SCSI
+  and IDE subsystems for the devices mentioned above.
+
+  The easiest way to implement a CAN device driver is as a character
+  device without such a (complete) abstraction layer, as is done by most
+  existing drivers.  The right way, however, would be to add such a
+  layer with all the functionality like registering for certain CAN
+  IDs, supporting several open file descriptors and (de)multiplexing
+  CAN frames between them, (sophisticated) queueing of CAN frames, and
+  providing an API for device drivers to register with.  However, then
+  it would be no more difficult, or may be even easier, to use the
+  networking framework provided by the Linux kernel, and this is what
+  Socket CAN does.
+
+  The use of the networking framework of the Linux kernel is just the
+  natural and most appropriate way to implement CAN for Linux.
+
+3. Socket CAN concept
+---------------------
+
+  As described in chapter 2 it is the main goal of Socket CAN to
+  provide a socket interface to user space applications which builds
+  upon the Linux network layer. In contrast to the commonly known
+  TCP/IP and ethernet networking, the CAN bus is a broadcast-only(!)
+  medium that has no MAC-layer addressing like ethernet. The CAN-identifier
+  (can_id) is used for arbitration on the CAN-bus. Therefore the CAN-IDs
+  have to be chosen uniquely on the bus. When designing a CAN-ECU
+  network the CAN-IDs are mapped to be sent by a specific ECU.
+  For this reason a CAN-ID can be treated best as a kind of source address.
+
+  3.1 receive lists
+
+  The network transparent access of multiple applications leads to the
+  problem that different applications may be interested in the same
+  CAN-IDs from the same CAN network interface. The Socket CAN core
+  module - which implements the protocol family CAN - provides several
+  high efficient receive lists for this reason. If e.g. a user space
+  application opens a CAN RAW socket, the raw protocol module itself
+  requests the (range of) CAN-IDs from the Socket CAN core that are
+  requested by the user. The subscription and unsubscription of
+  CAN-IDs can be done for specific CAN interfaces or for all(!) known
+  CAN interfaces with the can_rx_(un)register() functions provided to
+  CAN protocol modules by the SocketCAN core (see chapter 5).
+  To optimize the CPU usage at runtime the receive lists are split up
+  into several specific lists per device that match the requested
+  filter complexity for a given use-case.
+
+  3.2 local loopback of sent frames
+
+  As known from other networking concepts the data exchanging
+  applications may run on the same or different nodes without any
+  change (except for the according addressing information):
+
+         ___   ___   ___                   _______   ___
+        | _ | | _ | | _ |                 | _   _ | | _ |
+        ||A|| ||B|| ||C||                 ||A| |B|| ||C||
+        |___| |___| |___|                 |_______| |___|
+          |     |     |                       |       |
+        -----------------(1)- CAN bus -(2)---------------
+
+  To ensure that application A receives the same information in the
+  example (2) as it would receive in example (1) there is need for
+  some kind of local loopback of the sent CAN frames on the appropriate
+  node.
+
+  The Linux network devices (by default) just can handle the
+  transmission and reception of media dependent frames. Due to the
+  arbritration on the CAN bus the transmission of a low prio CAN-ID
+  may be delayed by the reception of a high prio CAN frame. To
+  reflect the correct* traffic on the node the loopback of the sent
+  data has to be performed right after a successful transmission. If
+  the CAN network interface is not capable of performing the loopback for
+  some reason the SocketCAN core can do this task as a fallback solution.
+  See chapter 6.2 for details (recommended).
+
+  The loopback functionality is enabled by default to reflect standard
+  networking behaviour for CAN applications. Due to some requests from
+  the RT-SocketCAN group the loopback optionally may be disabled for each
+  separate socket. See sockopts from the CAN RAW sockets in chapter 4.1.
+
+  * = you really like to have this when you're running analyser tools
+      like 'candump' or 'cansniffer' on the (same) node.
+
+  3.3 network security issues (capabilities)
+
+  The Controller Area Network is a local field bus transmitting only
+  broadcast messages without any routing and security concepts.
+  In the majority of cases the user application has to deal with
+  raw CAN frames. Therefore it might be reasonable NOT to restrict
+  the CAN access only to the user root, as known from other networks.
+  Since the currently implemented CAN_RAW and CAN_BCM sockets can only
+  send and receive frames to/from CAN interfaces it does not affect
+  security of others networks to allow all users to access the CAN.
+  To enable non-root users to access CAN_RAW and CAN_BCM protocol
+  sockets the Kconfig options CAN_RAW_USER and/or CAN_BCM_USER may be
+  selected at kernel compile time.
+
+  3.4 network problem notifications
+
+  The use of the CAN bus may lead to several problems on the physical
+  and media access control layer. Detecting and logging of these lower
+  layer problems is a vital requirement for CAN users to identify
+  hardware issues on the physical transceiver layer as well as
+  arbitration problems and error frames caused by the different
+  ECUs. The occurrence of detected errors are important for diagnosis
+  and have to be logged together with the exact timestamp. For this
+  reason the CAN interface driver can generate so called Error Frames
+  that can optionally be passed to the user application in the same
+  way as other CAN frames. Whenever an error on the physical layer
+  or the MAC layer is detected (e.g. by the CAN controller) the driver
+  creates an appropriate error frame. Error frames can be requested by
+  the user application using the common CAN filter mechanisms. Inside
+  this filter definition the (interested) type of errors may be
+  selected. The reception of error frames is disabled by default.
+
+4. How to use Socket CAN
+------------------------
+
+  Like TCP/IP, you first need to open a socket for communicating over a
+  CAN network. Since Socket CAN implements a new protocol family, you
+  need to pass PF_CAN as the first argument to the socket(2) system
+  call. Currently, there are two CAN protocols to choose from, the raw
+  socket protocol and the broadcast manager (BCM). So to open a socket,
+  you would write
+
+    s = socket(PF_CAN, SOCK_RAW, CAN_RAW);
+
+  and
+
+    s = socket(PF_CAN, SOCK_DGRAM, CAN_BCM);
+
+  respectively.  After the successful creation of the socket, you would
+  normally use the bind(2) system call to bind the socket to a CAN
+  interface (which is different from TCP/IP due to different addressing
+  - see chapter 3). After binding (CAN_RAW) or connecting (CAN_BCM)
+  the socket, you can read(2) and write(2) from/to the socket or use
+  send(2), sendto(2), sendmsg(2) and the recv* counterpart operations
+  on the socket as usual. There are also CAN specific socket options
+  described below.
+
+  The basic CAN frame structure and the sockaddr structure are defined
+  in include/linux/can.h:
+
+    struct can_frame {
+            canid_t can_id;  /* 32 bit CAN_ID + EFF/RTR/ERR flags */
+            __u8    can_dlc; /* data length code: 0 .. 8 */
+            __u8    data[8] __attribute__((aligned(8)));
+    };
+
+  The alignment of the (linear) payload data[] to a 64bit boundary
+  allows the user to define own structs and unions to easily access the
+  CAN payload. There is no given byteorder on the CAN bus by
+  default. A read(2) system call on a CAN_RAW socket transfers a
+  struct can_frame to the user space.
+
+  The sockaddr_can structure has an interface index like the
+  PF_PACKET socket, that also binds to a specific interface:
+
+    struct sockaddr_can {
+            sa_family_t can_family;
+            int         can_ifindex;
+            union {
+                    struct { canid_t rx_id, tx_id; } tp16;
+                    struct { canid_t rx_id, tx_id; } tp20;
+                    struct { canid_t rx_id, tx_id; } mcnet;
+                    struct { canid_t rx_id, tx_id; } isotp;
+            } can_addr;
+    };
+
+  To determine the interface index an appropriate ioctl() has to
+  be used (example for CAN_RAW sockets without error checking):
+
+    int s;
+    struct sockaddr_can addr;
+    struct ifreq ifr;
+
+    s = socket(PF_CAN, SOCK_RAW, CAN_RAW);
+
+    strcpy(ifr.ifr_name, "can0" );
+    ioctl(s, SIOCGIFINDEX, &ifr);
+
+    addr.can_family = AF_CAN;
+    addr.can_ifindex = ifr.ifr_ifindex;
+
+    bind(s, (struct sockaddr *)&addr, sizeof(addr));
+
+    (..)
+
+  To bind a socket to all(!) CAN interfaces the interface index must
+  be 0 (zero). In this case the socket receives CAN frames from every
+  enabled CAN interface. To determine the originating CAN interface
+  the system call recvfrom(2) may be used instead of read(2). To send
+  on a socket that is bound to 'any' interface sendto(2) is needed to
+  specify the outgoing interface.
+
+  Reading CAN frames from a bound CAN_RAW socket (see above) consists
+  of reading a struct can_frame:
+
+    struct can_frame frame;
+
+    nbytes = read(s, &frame, sizeof(struct can_frame));
+
+    if (nbytes < 0) {
+            perror("can raw socket read");
+            return 1;
+    }
+
+    /* paraniod check ... */
+    if (nbytes < sizeof(struct can_frame)) {
+            fprintf(stderr, "read: incomplete CAN frame\n");
+            return 1;
+    }
+
+    /* do something with the received CAN frame */
+
+  Writing CAN frames can be done similarly, with the write(2) system call:
+
+    nbytes = write(s, &frame, sizeof(struct can_frame));
+
+  When the CAN interface is bound to 'any' existing CAN interface
+  (addr.can_ifindex = 0) it is recommended to use recvfrom(2) if the
+  information about the originating CAN interface is needed:
+
+    struct sockaddr_can addr;
+    struct ifreq ifr;
+    socklen_t len = sizeof(addr);
+    struct can_frame frame;
+
+    nbytes = recvfrom(s, &frame, sizeof(struct can_frame),
+                      0, (struct sockaddr*)&addr, &len);
+
+    /* get interface name of the received CAN frame */
+    ifr.ifr_ifindex = addr.can_ifindex;
+    ioctl(s, SIOCGIFNAME, &ifr);
+    printf("Received a CAN frame from interface %s", ifr.ifr_name);
+
+  To write CAN frames on sockets bound to 'any' CAN interface the
+  outgoing interface has to be defined certainly.
+
+    strcpy(ifr.ifr_name, "can0");
+    ioctl(s, SIOCGIFINDEX, &ifr);
+    addr.can_ifindex = ifr.ifr_ifindex;
+    addr.can_family  = AF_CAN;
+
+    nbytes = sendto(s, &frame, sizeof(struct can_frame),
+                    0, (struct sockaddr*)&addr, sizeof(addr));
+
+  4.1 RAW protocol sockets with can_filters (SOCK_RAW)
+
+  Using CAN_RAW sockets is extensively comparable to the commonly
+  known access to CAN character devices. To meet the new possibilities
+  provided by the multi user SocketCAN approach, some reasonable
+  defaults are set at RAW socket binding time:
+
+  - The filters are set to exactly one filter receiving everything
+  - The socket only receives valid data frames (=> no error frames)
+  - The loopback of sent CAN frames is enabled (see chapter 3.2)
+  - The socket does not receive its own sent frames (in loopback mode)
+
+  These default settings may be changed before or after binding the socket.
+  To use the referenced definitions of the socket options for CAN_RAW
+  sockets, include <linux/can/raw.h>.
+
+  4.1.1 RAW socket option CAN_RAW_FILTER
+
+  The reception of CAN frames using CAN_RAW sockets can be controlled
+  by defining 0 .. n filters with the CAN_RAW_FILTER socket option.
+
+  The CAN filter structure is defined in include/linux/can.h:
+
+    struct can_filter {
+            canid_t can_id;
+            canid_t can_mask;
+    };
+
+  A filter matches, when
+
+    <received_can_id> & mask == can_id & mask
+
+  which is analogous to known CAN controllers hardware filter semantics.
+  The filter can be inverted in this semantic, when the CAN_INV_FILTER
+  bit is set in can_id element of the can_filter structure. In
+  contrast to CAN controller hardware filters the user may set 0 .. n
+  receive filters for each open socket separately:
+
+    struct can_filter rfilter[2];
+
+    rfilter[0].can_id   = 0x123;
+    rfilter[0].can_mask = CAN_SFF_MASK;
+    rfilter[1].can_id   = 0x200;
+    rfilter[1].can_mask = 0x700;
+
+    setsockopt(s, SOL_CAN_RAW, CAN_RAW_FILTER, &rfilter, sizeof(rfilter));
+
+  To disable the reception of CAN frames on the selected CAN_RAW socket:
+
+    setsockopt(s, SOL_CAN_RAW, CAN_RAW_FILTER, NULL, 0);
+
+  To set the filters to zero filters is quite obsolete as not read
+  data causes the raw socket to discard the received CAN frames. But
+  having this 'send only' use-case we may remove the receive list in the
+  Kernel to save a little (really a very little!) CPU usage.
+
+  4.1.2 RAW socket option CAN_RAW_ERR_FILTER
+
+  As described in chapter 3.4 the CAN interface driver can generate so
+  called Error Frames that can optionally be passed to the user
+  application in the same way as other CAN frames. The possible
+  errors are divided into different error classes that may be filtered
+  using the appropriate error mask. To register for every possible
+  error condition CAN_ERR_MASK can be used as value for the error mask.
+  The values for the error mask are defined in linux/can/error.h .
+
+    can_err_mask_t err_mask = ( CAN_ERR_TX_TIMEOUT | CAN_ERR_BUSOFF );
+
+    setsockopt(s, SOL_CAN_RAW, CAN_RAW_ERR_FILTER,
+               &err_mask, sizeof(err_mask));
+
+  4.1.3 RAW socket option CAN_RAW_LOOPBACK
+
+  To meet multi user needs the local loopback is enabled by default
+  (see chapter 3.2 for details). But in some embedded use-cases
+  (e.g. when only one application uses the CAN bus) this loopback
+  functionality can be disabled (separately for each socket):
+
+    int loopback = 0; /* 0 = disabled, 1 = enabled (default) */
+
+    setsockopt(s, SOL_CAN_RAW, CAN_RAW_LOOPBACK, &loopback, sizeof(loopback));
+
+  4.1.4 RAW socket option CAN_RAW_RECV_OWN_MSGS
+
+  When the local loopback is enabled, all the sent CAN frames are
+  looped back to the open CAN sockets that registered for the CAN
+  frames' CAN-ID on this given interface to meet the multi user
+  needs. The reception of the CAN frames on the same socket that was
+  sending the CAN frame is assumed to be unwanted and therefore
+  disabled by default. This default behaviour may be changed on
+  demand:
+
+    int recv_own_msgs = 1; /* 0 = disabled (default), 1 = enabled */
+
+    setsockopt(s, SOL_CAN_RAW, CAN_RAW_RECV_OWN_MSGS,
+               &recv_own_msgs, sizeof(recv_own_msgs));
+
+  4.2 Broadcast Manager protocol sockets (SOCK_DGRAM)
+  4.3 connected transport protocols (SOCK_SEQPACKET)
+  4.4 unconnected transport protocols (SOCK_DGRAM)
+
+
+5. Socket CAN core module
+-------------------------
+
+  The Socket CAN core module implements the protocol family
+  PF_CAN. CAN protocol modules are loaded by the core module at
+  runtime. The core module provides an interface for CAN protocol
+  modules to subscribe needed CAN IDs (see chapter 3.1).
+
+  5.1 can.ko module params
+
+  - stats_timer: To calculate the Socket CAN core statistics
+    (e.g. current/maximum frames per second) this 1 second timer is
+    invoked at can.ko module start time by default. This timer can be
+    disabled by using stattimer=0 on the module comandline.
+
+  - debug: (removed since SocketCAN SVN r546)
+
+  5.2 procfs content
+
+  As described in chapter 3.1 the Socket CAN core uses several filter
+  lists to deliver received CAN frames to CAN protocol modules. These
+  receive lists, their filters and the count of filter matches can be
+  checked in the appropriate receive list. All entries contain the
+  device and a protocol module identifier:
+
+    foo@bar:~$ cat /proc/net/can/rcvlist_all
+
+    receive list 'rx_all':
+      (vcan3: no entry)
+      (vcan2: no entry)
+      (vcan1: no entry)
+      device   can_id   can_mask  function  userdata   matches  ident
+       vcan0     000    00000000  f88e6370  f6c6f400         0  raw
+      (any: no entry)
+
+  In this example an application requests any CAN traffic from vcan0.
+
+    rcvlist_all - list for unfiltered entries (no filter operations)
+    rcvlist_eff - list for single extended frame (EFF) entries
+    rcvlist_err - list for error frames masks
+    rcvlist_fil - list for mask/value filters
+    rcvlist_inv - list for mask/value filters (inverse semantic)
+    rcvlist_sff - list for single standard frame (SFF) entries
+
+  Additional procfs files in /proc/net/can
+
+    stats       - Socket CAN core statistics (rx/tx frames, match ratios, ...)
+    reset_stats - manual statistic reset
+    version     - prints the Socket CAN core version and the ABI version
+
+  5.3 writing own CAN protocol modules
+
+  To implement a new protocol in the protocol family PF_CAN a new
+  protocol has to be defined in include/linux/can.h .
+  The prototypes and definitions to use the Socket CAN core can be
+  accessed by including include/linux/can/core.h .
+  In addition to functions that register the CAN protocol and the
+  CAN device notifier chain there are functions to subscribe CAN
+  frames received by CAN interfaces and to send CAN frames:
+
+    can_rx_register   - subscribe CAN frames from a specific interface
+    can_rx_unregister - unsubscribe CAN frames from a specific interface
+    can_send          - transmit a CAN frame (optional with local loopback)
+
+  For details see the kerneldoc documentation in net/can/af_can.c or
+  the source code of net/can/raw.c or net/can/bcm.c .
+
+6. CAN network drivers
+----------------------
+
+  Writing a CAN network device driver is much easier than writing a
+  CAN character device driver. Similar to other known network device
+  drivers you mainly have to deal with:
+
+  - TX: Put the CAN frame from the socket buffer to the CAN controller.
+  - RX: Put the CAN frame from the CAN controller to the socket buffer.
+
+  See e.g. at Documentation/networking/netdevices.txt . The differences
+  for writing CAN network device driver are described below:
+
+  6.1 general settings
+
+    dev->type  = ARPHRD_CAN; /* the netdevice hardware type */
+    dev->flags = IFF_NOARP;  /* CAN has no arp */
+
+    dev->mtu   = sizeof(struct can_frame);
+
+  The struct can_frame is the payload of each socket buffer in the
+  protocol family PF_CAN.
+
+  6.2 local loopback of sent frames
+
+  As described in chapter 3.2 the CAN network device driver should
+  support a local loopback functionality similar to the local echo
+  e.g. of tty devices. In this case the driver flag IFF_ECHO has to be
+  set to prevent the PF_CAN core from locally echoing sent frames
+  (aka loopback) as fallback solution:
+
+    dev->flags = (IFF_NOARP | IFF_ECHO);
+
+  6.3 CAN controller hardware filters
+
+  To reduce the interrupt load on deep embedded systems some CAN
+  controllers support the filtering of CAN IDs or ranges of CAN IDs.
+  These hardware filter capabilities vary from controller to
+  controller and have to be identified as not feasible in a multi-user
+  networking approach. The use of the very controller specific
+  hardware filters could make sense in a very dedicated use-case, as a
+  filter on driver level would affect all users in the multi-user
+  system. The high efficient filter sets inside the PF_CAN core allow
+  to set different multiple filters for each socket separately.
+  Therefore the use of hardware filters goes to the category 'handmade
+  tuning on deep embedded systems'. The author is running a MPC603e
+  @133MHz with four SJA1000 CAN controllers from 2002 under heavy bus
+  load without any problems ...
+
+  6.4 currently supported CAN hardware (September 2007)
+
+  On the project website http://developer.berlios.de/projects/socketcan
+  there are different drivers available:
+
+    vcan:    Virtual CAN interface driver (if no real hardware is available)
+    sja1000: Philips SJA1000 CAN controller (recommended)
+    i82527:  Intel i82527 CAN controller
+    mscan:   Motorola/Freescale CAN controller (e.g. inside SOC MPC5200)
+    ccan:    CCAN controller core (e.g. inside SOC h7202)
+    slcan:   For a bunch of CAN adaptors that are attached via a
+             serial line ASCII protocol (for serial / USB adaptors)
+
+  Additionally the different CAN adaptors (ISA/PCI/PCMCIA/USB/Parport)
+  from PEAK Systemtechnik support the CAN netdevice driver model
+  since Linux driver v6.0: http://www.peak-system.com/linux/index.htm
+
+  Please check the Mailing Lists on the berlios OSS project website.
+
+  6.5 todo (September 2007)
+
+  The configuration interface for CAN network drivers is still an open
+  issue that has not been finalized in the socketcan project. Also the
+  idea of having a library module (candev.ko) that holds functions
+  that are needed by all CAN netdevices is not ready to ship.
+  Your contribution is welcome.
+
+7. Credits
+----------
+
+  Oliver Hartkopp (PF_CAN core, filters, drivers, bcm)
+  Urs Thuermann (PF_CAN core, kernel integration, socket interfaces, raw, vcan)
+  Jan Kizka (RT-SocketCAN core, Socket-API reconciliation)
+  Wolfgang Grandegger (RT-SocketCAN core & drivers, Raw Socket-API reviews)
+  Robert Schwebel (design reviews, PTXdist integration)
+  Marc Kleine-Budde (design reviews, Kernel 2.6 cleanups, drivers)
+  Benedikt Spranger (reviews)
+  Thomas Gleixner (LKML reviews, coding style, posting hints)
+  Andrey Volkov (kernel subtree structure, ioctls, mscan driver)
+  Matthias Brukner (first SJA1000 CAN netdevice implementation Q2/2003)
+  Klaus Hitschler (PEAK driver integration)
+  Uwe Koppe (CAN netdevices with PF_PACKET approach)
+  Michael Schulze (driver layer loopback requirement, RT CAN drivers review)
index afb66f9..39131a3 100644 (file)
@@ -14,24 +14,35 @@ Introduction
 ============
 
 Datagram Congestion Control Protocol (DCCP) is an unreliable, connection
-based protocol designed to solve issues present in UDP and TCP particularly
-for real time and multimedia traffic.
+oriented protocol designed to solve issues present in UDP and TCP, particularly
+for real-time and multimedia (streaming) traffic.
+It divides into a base protocol (RFC 4340) and plugable congestion control
+modules called CCIDs. Like plugable TCP congestion control, at least one CCID
+needs to be enabled in order for the protocol to function properly. In the Linux
+implementation, this is the TCP-like CCID2 (RFC 4341). Additional CCIDs, such as
+the TCP-friendly CCID3 (RFC 4342), are optional.
+For a brief introduction to CCIDs and suggestions for choosing a CCID to match
+given applications, see section 10 of RFC 4340.
 
 It has a base protocol and pluggable congestion control IDs (CCIDs).
 
-It is at proposed standard RFC status and the homepage for DCCP as a protocol
-is at:
-       http://www.read.cs.ucla.edu/dccp/
+DCCP is a Proposed Standard (RFC 2026), and the homepage for DCCP as a protocol
+is at http://www.ietf.org/html.charters/dccp-charter.html
 
 Missing features
 ================
 
-The DCCP implementation does not currently have all the features that are in
-the RFC.
+The Linux DCCP implementation does not currently support all the features that are
+specified in RFCs 4340...42.
 
 The known bugs are at:
        http://linux-net.osdl.org/index.php/TODO#DCCP
 
+For more up-to-date versions of the DCCP implementation, please consider using
+the experimental DCCP test tree; instructions for checking this out are on:
+http://linux-net.osdl.org/index.php/DCCP_Testing#Experimental_DCCP_source_tree
+
+
 Socket options
 ==============
 
@@ -46,6 +57,12 @@ can be set before calling bind().
 DCCP_SOCKOPT_GET_CUR_MPS is read-only and retrieves the current maximum packet
 size (application payload size) in bytes, see RFC 4340, section 14.
 
+DCCP_SOCKOPT_SERVER_TIMEWAIT enables the server (listening socket) to hold
+timewait state when closing the connection (RFC 4340, 8.3). The usual case is
+that the closing server sends a CloseReq, whereupon the client holds timewait
+state. When this boolean socket option is on, the server sends a Close instead
+and will enter TIMEWAIT. This option must be set after accept() returns.
+
 DCCP_SOCKOPT_SEND_CSCOV and DCCP_SOCKOPT_RECV_CSCOV are used for setting the
 partial checksum coverage (RFC 4340, sec. 9.2). The default is that checksums
 always cover the entire packet and that only fully covered application data is
@@ -72,6 +89,8 @@ DCCP_SOCKOPT_CCID_TX_INFO
        Returns a `struct tfrc_tx_info' in optval; the buffer for optval and
        optlen must be set to at least sizeof(struct tfrc_tx_info).
 
+On unidirectional connections it is useful to close the unused half-connection
+via shutdown (SHUT_WR or SHUT_RD): this will reduce per-packet processing costs.
 
 Sysctl variables
 ================
@@ -123,6 +142,12 @@ sync_ratelimit = 125 ms
        sequence-invalid packets on the same socket (RFC 4340, 7.5.4). The unit
        of this parameter is milliseconds; a value of 0 disables rate-limiting.
 
+IOCTLS
+======
+FIONREAD
+       Works as in udp(7): returns in the `int' argument pointer the size of
+       the next pending datagram in bytes, or 0 when no datagram is pending.
+
 Notes
 =====
 
index 6f7872b..17a6e46 100644 (file)
@@ -446,6 +446,33 @@ tcp_dma_copybreak - INTEGER
        and CONFIG_NET_DMA is enabled.
        Default: 4096
 
+UDP variables:
+
+udp_mem - vector of 3 INTEGERs: min, pressure, max
+       Number of pages allowed for queueing by all UDP sockets.
+
+       min: Below this number of pages UDP is not bothered about its
+       memory appetite. When amount of memory allocated by UDP exceeds
+       this number, UDP starts to moderate memory usage.
+
+       pressure: This value was introduced to follow format of tcp_mem.
+
+       max: Number of pages allowed for queueing by all UDP sockets.
+
+       Default is calculated at boot time from amount of available memory.
+
+udp_rmem_min - INTEGER
+       Minimal size of receive buffer used by UDP sockets in moderation.
+       Each UDP socket is able to use the size for receiving data, even if
+       total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
+       Default: 4096
+
+udp_wmem_min - INTEGER
+       Minimal size of send buffer used by UDP sockets in moderation.
+       Each UDP socket is able to use the size for sending data, even if
+       total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
+       Default: 4096
+
 CIPSOv4 Variables:
 
 cipso_cache_enable - BOOLEAN
diff --git a/Documentation/networking/shaper.txt b/Documentation/networking/shaper.txt
deleted file mode 100644 (file)
index 6c4ebb6..0000000
+++ /dev/null
@@ -1,48 +0,0 @@
-Traffic Shaper For Linux
-
-This is the current BETA release of the traffic shaper for Linux. It works
-within the following limits:
-
-o      Minimum shaping speed is currently about 9600 baud (it can only
-shape down to 1 byte per clock tick)
-
-o      Maximum is about 256K, it will go above this but get a bit blocky.
-
-o      If you ifconfig the master device that a shaper is attached to down
-then your machine will follow.
-
-o      The shaper must be a module.
-
-
-Setup:
-
-       A shaper device is configured using the shapeconfig program.
-Typically you will do something like this
-
-shapecfg attach shaper0 eth1
-shapecfg speed shaper0 64000
-ifconfig shaper0 myhost netmask 255.255.255.240 broadcast 1.2.3.4.255 up
-route add -net some.network netmask a.b.c.d dev shaper0
-
-The shaper should have the same IP address as the device it is attached to
-for normal use.
-
-Gotchas:
-
-       The shaper shapes transmitted traffic. It's rather impossible to
-shape received traffic except at the end (or a router) transmitting it.
-
-       Gated/routed/rwhod/mrouted all see the shaper as an additional device
-and will treat it as such unless patched. Note that for mrouted you can run
-mrouted tunnels via a traffic shaper to control bandwidth usage.
-
-       The shaper is device/route based. This makes it very easy to use
-with any setup BUT less flexible. You may need to use iproute2 to set up
-multiple route tables to get the flexibility.
-
-       There is no "borrowing" or "sharing" scheme. This is a simple
-traffic limiter. We implement Van Jacobson and Sally Floyd's CBQ
-architecture into Linux 2.2. This is the preferred solution. Shaper is
-for simple or back compatible setups.
-
-Alan
index b6409ca..3870f28 100644 (file)
 
   This displays UDP-Lite statistics variables, whose meaning is as follows.
 
-   InDatagrams:     Total number of received datagrams.
+   InDatagrams:     The total number of datagrams delivered to users.
 
    NoPorts:         Number of packets received to an unknown port.
                     These cases are counted separately (not as InErrors).
diff --git a/Documentation/networking/xfrm_proc.txt b/Documentation/networking/xfrm_proc.txt
new file mode 100644 (file)
index 0000000..53c1a58
--- /dev/null
@@ -0,0 +1,70 @@
+XFRM proc - /proc/net/xfrm_* files
+==================================
+Masahide NAKAMURA <nakam@linux-ipv6.org>
+
+
+Transformation Statistics
+-------------------------
+xfrm_proc is a statistics shown factor dropped by transformation
+for developer.
+It is a counter designed from current transformation source code
+and defined like linux private MIB.
+
+Inbound statistics
+~~~~~~~~~~~~~~~~~~
+XfrmInError:
+       All errors which is not matched others
+XfrmInBufferError:
+       No buffer is left
+XfrmInHdrError:
+       Header error
+XfrmInNoStates:
+       No state is found
+       i.e. Either inbound SPI, address, or IPsec protocol at SA is wrong
+XfrmInStateProtoError:
+       Transformation protocol specific error
+       e.g. SA key is wrong
+XfrmInStateModeError:
+       Transformation mode specific error
+XfrmInSeqOutOfWindow:
+       Sequence out of window
+XfrmInStateExpired:
+       State is expired
+XfrmInStateMismatch:
+       State has mismatch option
+       e.g. UDP encapsulation type is mismatch
+XfrmInStateInvalid:
+       State is invalid
+XfrmInTmplMismatch:
+       No matching template for states
+       e.g. Inbound SAs are correct but SP rule is wrong
+XfrmInNoPols:
+       No policy is found for states
+       e.g. Inbound SAs are correct but no SP is found
+XfrmInPolBlock:
+       Policy discards
+XfrmInPolError:
+       Policy error
+
+Outbound errors
+~~~~~~~~~~~~~~~
+XfrmOutError:
+       All errors which is not matched others
+XfrmOutBundleGenError:
+       Bundle generation error
+XfrmOutBundleCheckError:
+       Bundle check error
+XfrmOutNoStates:
+       No state is found
+XfrmOutStateProtoError:
+       Transformation protocol specific error
+XfrmOutStateModeError:
+       Transformation mode specific error
+XfrmOutStateExpired:
+       State is expired
+XfrmOutPolBlock:
+       Policy discards
+XfrmOutPolDead:
+       Policy is dead
+XfrmOutPolError:
+       Policy error
index 2937122..ba05e80 100644 (file)
@@ -646,6 +646,17 @@ M: ecashin@coraid.com
 W:     http://www.coraid.com/support/linux
 S:     Supported
 
+ATHEROS ATH5K WIRELESS DRIVER
+P:     Jiri Slaby
+M:     jirislaby@gmail.com
+P:     Nick Kossifidis
+M:     mickflemm@gmail.com
+P:     Luis R. Rodriguez
+M:     mcgrof@gmail.com
+L:     linux-wireless@vger.kernel.org
+L:     ath5k-devel@lists.ath5k.org
+S:     Maintained
+
 ATL1 ETHERNET DRIVER
 P:     Jay Cliburn
 M:     jcliburn@gmail.com
@@ -809,7 +820,7 @@ P:  Stefano Brivio
 M:     stefano.brivio@polimi.it
 L:     linux-wireless@vger.kernel.org
 W:     http://bcm43xx.berlios.de/
-S:     Maintained
+S:     Obsolete
 
 BEFS FILE SYSTEM
 P:     Sergey S. Kostyliov
@@ -982,6 +993,15 @@ M: corbet@lwn.net
 L:     video4linux-list@redhat.com
 S:     Maintained
 
+CAN NETWORK LAYER
+P:     Urs Thuermann
+M:     urs.thuermann@volkswagen.de
+P:     Oliver Hartkopp
+M:     oliver.hartkopp@volkswagen.de
+L:     socketcan-core@lists.berlios.de
+W:     http://developer.berlios.de/projects/socketcan/
+S:     Maintained
+
 CALGARY x86-64 IOMMU
 P:     Muli Ben-Yehuda
 M:     muli@il.ibm.com
@@ -2027,10 +2047,12 @@ W:      http://sourceforge.net/projects/e1000/
 S:     Supported
 
 INTEL PRO/WIRELESS 2100 NETWORK CONNECTION SUPPORT
-P:     Yi Zhu
+P:     Zhu Yi
 M:     yi.zhu@intel.com
 P:     James Ketrenos
 M:     jketreno@linux.intel.com
+P:     Reinette Chatre
+M:     reinette.chatre@intel.com
 L:     linux-wireless@vger.kernel.org
 L:     ipw2100-devel@lists.sourceforge.net
 W:     http://lists.sourceforge.net/mailman/listinfo/ipw2100-devel
@@ -2038,10 +2060,12 @@ W:      http://ipw2100.sourceforge.net
 S:     Supported
 
 INTEL PRO/WIRELESS 2915ABG NETWORK CONNECTION SUPPORT
-P:     Yi Zhu
+P:     Zhu Yi
 M:     yi.zhu@intel.com
 P:     James Ketrenos
 M:     jketreno@linux.intel.com
+P:     Reinette Chatre
+M:     reinette.chatre@intel.com
 L:     linux-wireless@vger.kernel.org
 L:     ipw2100-devel@lists.sourceforge.net
 W:     http://lists.sourceforge.net/mailman/listinfo/ipw2100-devel
@@ -2051,6 +2075,8 @@ S:        Supported
 INTEL WIRELESS WIFI LINK (iwlwifi)
 P:     Zhu Yi
 M:     yi.zhu@intel.com
+P:     Reinette Chatre
+M:     reinette.chatre@intel.com
 L:     linux-wireless@vger.kernel.org
 L:     ipw3945-devel@lists.sourceforge.net
 W:     http://intellinuxwireless.org
@@ -2482,6 +2508,16 @@ W:       http://linuxwireless.org/
 T:     git kernel.org:/pub/scm/linux/kernel/git/linville/wireless-2.6.git
 S:     Maintained
 
+MAC80211 PID RATE CONTROL
+P:     Stefano Brivio
+M:     stefano.brivio@polimi.it
+P:     Mattias Nissler
+M:     mattias.nissler@gmx.de
+L:     linux-wireless@vger.kernel.org
+W:     http://linuxwireless.org/en/developers/Documentation/mac80211/RateControl/PID
+T:     git kernel.org:/pub/scm/linux/kernel/git/linville/wireless-2.6.git
+S:     Maintained
+
 MACVLAN DRIVER
 P:     Patrick McHardy
 M:     kaber@trash.net
@@ -3183,6 +3219,12 @@ M:       mporter@kernel.crashing.org
 L:     linux-kernel@vger.kernel.org
 S:     Maintained
 
+RDC R6040 FAST ETHERNET DRIVER
+P:     Florian Fainelli
+M:     florian.fainelli@telecomint.eu
+L:     netdev@vger.kernel.org
+S:     Maintained
+
 READ-COPY UPDATE (RCU)
 P:     Dipankar Sarma
 M:     dipankar@in.ibm.com
index 08b117e..9898feb 100644 (file)
@@ -497,11 +497,6 @@ simeth_interrupt(int irq, void *dev_id)
 {
        struct net_device *dev = dev_id;
 
-       if ( dev == NULL ) {
-               printk(KERN_WARNING "simeth: irq %d for unknown device\n", irq);
-               return IRQ_NONE;
-       }
-
        /*
         * very simple loop because we get interrupts only when receiving
         */
index f47fcac..e636daa 100644 (file)
@@ -1,4 +1,4 @@
-obj-y  += setup.o pci.o time.o idle.o powersave.o iommu.o
+obj-y  += setup.o pci.o time.o idle.o powersave.o iommu.o dma_lib.o
 obj-$(CONFIG_PPC_PASEMI_MDIO)  += gpio_mdio.o
 obj-$(CONFIG_ELECTRA_IDE) += electra_ide.o
 obj-$(CONFIG_PPC_PASEMI_CPUFREQ) += cpufreq.o
diff --git a/arch/powerpc/platforms/pasemi/dma_lib.c b/arch/powerpc/platforms/pasemi/dma_lib.c
new file mode 100644 (file)
index 0000000..c529d8d
--- /dev/null
@@ -0,0 +1,488 @@
+/*
+ * Copyright (C) 2006-2007 PA Semi, Inc
+ *
+ * Common functions for DMA access on PA Semi PWRficient
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program 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 this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/of.h>
+
+#include <asm/pasemi_dma.h>
+
+#define MAX_TXCH 64
+#define MAX_RXCH 64
+
+static struct pasdma_status *dma_status;
+
+static void __iomem *iob_regs;
+static void __iomem *mac_regs[6];
+static void __iomem *dma_regs;
+
+static int base_hw_irq;
+
+static int num_txch, num_rxch;
+
+static struct pci_dev *dma_pdev;
+
+/* Bitmaps to handle allocation of channels */
+
+static DECLARE_BITMAP(txch_free, MAX_TXCH);
+static DECLARE_BITMAP(rxch_free, MAX_RXCH);
+
+/* pasemi_read_iob_reg - read IOB register
+ * @reg: Register to read (offset into PCI CFG space)
+ */
+unsigned int pasemi_read_iob_reg(unsigned int reg)
+{
+       return in_le32(iob_regs+reg);
+}
+EXPORT_SYMBOL(pasemi_read_iob_reg);
+
+/* pasemi_write_iob_reg - write IOB register
+ * @reg: Register to write to (offset into PCI CFG space)
+ * @val: Value to write
+ */
+void pasemi_write_iob_reg(unsigned int reg, unsigned int val)
+{
+       out_le32(iob_regs+reg, val);
+}
+EXPORT_SYMBOL(pasemi_write_iob_reg);
+
+/* pasemi_read_mac_reg - read MAC register
+ * @intf: MAC interface
+ * @reg: Register to read (offset into PCI CFG space)
+ */
+unsigned int pasemi_read_mac_reg(int intf, unsigned int reg)
+{
+       return in_le32(mac_regs[intf]+reg);
+}
+EXPORT_SYMBOL(pasemi_read_mac_reg);
+
+/* pasemi_write_mac_reg - write MAC register
+ * @intf: MAC interface
+ * @reg: Register to write to (offset into PCI CFG space)
+ * @val: Value to write
+ */
+void pasemi_write_mac_reg(int intf, unsigned int reg, unsigned int val)
+{
+       out_le32(mac_regs[intf]+reg, val);
+}
+EXPORT_SYMBOL(pasemi_write_mac_reg);
+
+/* pasemi_read_dma_reg - read DMA register
+ * @reg: Register to read (offset into PCI CFG space)
+ */
+unsigned int pasemi_read_dma_reg(unsigned int reg)
+{
+       return in_le32(dma_regs+reg);
+}
+EXPORT_SYMBOL(pasemi_read_dma_reg);
+
+/* pasemi_write_dma_reg - write DMA register
+ * @reg: Register to write to (offset into PCI CFG space)
+ * @val: Value to write
+ */
+void pasemi_write_dma_reg(unsigned int reg, unsigned int val)
+{
+       out_le32(dma_regs+reg, val);
+}
+EXPORT_SYMBOL(pasemi_write_dma_reg);
+
+static int pasemi_alloc_tx_chan(enum pasemi_dmachan_type type)
+{
+       int bit;
+       int start, limit;
+
+       switch (type & (TXCHAN_EVT0|TXCHAN_EVT1)) {
+       case TXCHAN_EVT0:
+               start = 0;
+               limit = 10;
+               break;
+       case TXCHAN_EVT1:
+               start = 10;
+               limit = MAX_TXCH;
+               break;
+       default:
+               start = 0;
+               limit = MAX_TXCH;
+               break;
+       }
+retry:
+       bit = find_next_bit(txch_free, MAX_TXCH, start);
+       if (bit >= limit)
+               return -ENOSPC;
+       if (!test_and_clear_bit(bit, txch_free))
+               goto retry;
+
+       return bit;
+}
+
+static void pasemi_free_tx_chan(int chan)
+{
+       BUG_ON(test_bit(chan, txch_free));
+       set_bit(chan, txch_free);
+}
+
+static int pasemi_alloc_rx_chan(void)
+{
+       int bit;
+retry:
+       bit = find_first_bit(rxch_free, MAX_RXCH);
+       if (bit >= MAX_TXCH)
+               return -ENOSPC;
+       if (!test_and_clear_bit(bit, rxch_free))
+               goto retry;
+
+       return bit;
+}
+
+static void pasemi_free_rx_chan(int chan)
+{
+       BUG_ON(test_bit(chan, rxch_free));
+       set_bit(chan, rxch_free);
+}
+
+/* pasemi_dma_alloc_chan - Allocate a DMA channel
+ * @type: Type of channel to allocate
+ * @total_size: Total size of structure to allocate (to allow for more
+ *             room behind the structure to be used by the client)
+ * @offset: Offset in bytes from start of the total structure to the beginning
+ *         of struct pasemi_dmachan. Needed when struct pasemi_dmachan is
+ *         not the first member of the client structure.
+ *
+ * pasemi_dma_alloc_chan allocates a DMA channel for use by a client. The
+ * type argument specifies whether it's a RX or TX channel, and in the case
+ * of TX channels which group it needs to belong to (if any).
+ *
+ * Returns a pointer to the total structure allocated on success, NULL
+ * on failure.
+ */
+void *pasemi_dma_alloc_chan(enum pasemi_dmachan_type type,
+                           int total_size, int offset)
+{
+       void *buf;
+       struct pasemi_dmachan *chan;
+       int chno;
+
+       BUG_ON(total_size < sizeof(struct pasemi_dmachan));
+
+       buf = kzalloc(total_size, GFP_KERNEL);
+
+       if (!buf)
+               return NULL;
+       chan = buf + offset;
+
+       chan->priv = buf;
+
+       switch (type & (TXCHAN|RXCHAN)) {
+       case RXCHAN:
+               chno = pasemi_alloc_rx_chan();
+               chan->chno = chno;
+               chan->irq = irq_create_mapping(NULL,
+                                              base_hw_irq + num_txch + chno);
+               chan->status = &dma_status->rx_sta[chno];
+               break;
+       case TXCHAN:
+               chno = pasemi_alloc_tx_chan(type);
+               chan->chno = chno;
+               chan->irq = irq_create_mapping(NULL, base_hw_irq + chno);
+               chan->status = &dma_status->tx_sta[chno];
+               break;
+       }
+
+       chan->chan_type = type;
+
+       return chan;
+}
+EXPORT_SYMBOL(pasemi_dma_alloc_chan);
+
+/* pasemi_dma_free_chan - Free a previously allocated channel
+ * @chan: Channel to free
+ *
+ * Frees a previously allocated channel. It will also deallocate any
+ * descriptor ring associated with the channel, if allocated.
+ */
+void pasemi_dma_free_chan(struct pasemi_dmachan *chan)
+{
+       if (chan->ring_virt)
+               pasemi_dma_free_ring(chan);
+
+       switch (chan->chan_type & (RXCHAN|TXCHAN)) {
+       case RXCHAN:
+               pasemi_free_rx_chan(chan->chno);
+               break;
+       case TXCHAN:
+               pasemi_free_tx_chan(chan->chno);
+               break;
+       }
+
+       kfree(chan->priv);
+}
+EXPORT_SYMBOL(pasemi_dma_free_chan);
+
+/* pasemi_dma_alloc_ring - Allocate descriptor ring for a channel
+ * @chan: Channel for which to allocate
+ * @ring_size: Ring size in 64-bit (8-byte) words
+ *
+ * Allocate a descriptor ring for a channel. Returns 0 on success, errno
+ * on failure. The passed in struct pasemi_dmachan is updated with the
+ * virtual and DMA addresses of the ring.
+ */
+int pasemi_dma_alloc_ring(struct pasemi_dmachan *chan, int ring_size)
+{
+       BUG_ON(chan->ring_virt);
+
+       chan->ring_size = ring_size;
+
+       chan->ring_virt = dma_alloc_coherent(&dma_pdev->dev,
+                                            ring_size * sizeof(u64),
+                                            &chan->ring_dma, GFP_KERNEL);
+
+       if (!chan->ring_virt)
+               return -ENOMEM;
+
+       memset(chan->ring_virt, 0, ring_size * sizeof(u64));
+
+       return 0;
+}
+EXPORT_SYMBOL(pasemi_dma_alloc_ring);
+
+/* pasemi_dma_free_ring - Free an allocated descriptor ring for a channel
+ * @chan: Channel for which to free the descriptor ring
+ *
+ * Frees a previously allocated descriptor ring for a channel.
+ */
+void pasemi_dma_free_ring(struct pasemi_dmachan *chan)
+{
+       BUG_ON(!chan->ring_virt);
+
+       dma_free_coherent(&dma_pdev->dev, chan->ring_size * sizeof(u64),
+                         chan->ring_virt, chan->ring_dma);
+       chan->ring_virt = NULL;
+       chan->ring_size = 0;
+       chan->ring_dma = 0;
+}
+EXPORT_SYMBOL(pasemi_dma_free_ring);
+
+/* pasemi_dma_start_chan - Start a DMA channel
+ * @chan: Channel to start
+ * @cmdsta: Additional CCMDSTA/TCMDSTA bits to write
+ *
+ * Enables (starts) a DMA channel with optional additional arguments.
+ */
+void pasemi_dma_start_chan(const struct pasemi_dmachan *chan, const u32 cmdsta)
+{
+       if (chan->chan_type == RXCHAN)
+               pasemi_write_dma_reg(PAS_DMA_RXCHAN_CCMDSTA(chan->chno),
+                                    cmdsta | PAS_DMA_RXCHAN_CCMDSTA_EN);
+       else
+               pasemi_write_dma_reg(PAS_DMA_TXCHAN_TCMDSTA(chan->chno),
+                                    cmdsta | PAS_DMA_TXCHAN_TCMDSTA_EN);
+}
+EXPORT_SYMBOL(pasemi_dma_start_chan);
+
+/* pasemi_dma_stop_chan - Stop a DMA channel
+ * @chan: Channel to stop
+ *
+ * Stops (disables) a DMA channel. This is done by setting the ST bit in the
+ * CMDSTA register and waiting on the ACT (active) bit to clear, then
+ * finally disabling the whole channel.
+ *
+ * This function will only try for a short while for the channel to stop, if
+ * it doesn't it will return failure.
+ *
+ * Returns 1 on success, 0 on failure.
+ */
+#define MAX_RETRIES 5000
+int pasemi_dma_stop_chan(const struct pasemi_dmachan *chan)
+{
+       int reg, retries;
+       u32 sta;
+
+       if (chan->chan_type == RXCHAN) {
+               reg = PAS_DMA_RXCHAN_CCMDSTA(chan->chno);
+               pasemi_write_dma_reg(reg, PAS_DMA_RXCHAN_CCMDSTA_ST);
+               for (retries = 0; retries < MAX_RETRIES; retries++) {
+                       sta = pasemi_read_dma_reg(reg);
+                       if (!(sta & PAS_DMA_RXCHAN_CCMDSTA_ACT)) {
+                               pasemi_write_dma_reg(reg, 0);
+                               return 1;
+                       }
+                       cond_resched();
+               }
+       } else {
+               reg = PAS_DMA_TXCHAN_TCMDSTA(chan->chno);
+               pasemi_write_dma_reg(reg, PAS_DMA_TXCHAN_TCMDSTA_ST);
+               for (retries = 0; retries < MAX_RETRIES; retries++) {
+                       sta = pasemi_read_dma_reg(reg);
+                       if (!(sta & PAS_DMA_TXCHAN_TCMDSTA_ACT)) {
+                               pasemi_write_dma_reg(reg, 0);
+                               return 1;
+                       }
+                       cond_resched();
+               }
+       }
+
+       return 0;
+}
+EXPORT_SYMBOL(pasemi_dma_stop_chan);
+
+/* pasemi_dma_alloc_buf - Allocate a buffer to use for DMA
+ * @chan: Channel to allocate for
+ * @size: Size of buffer in bytes
+ * @handle: DMA handle
+ *
+ * Allocate a buffer to be used by the DMA engine for read/write,
+ * similar to dma_alloc_coherent().
+ *
+ * Returns the virtual address of the buffer, or NULL in case of failure.
+ */
+void *pasemi_dma_alloc_buf(struct pasemi_dmachan *chan, int size,
+                          dma_addr_t *handle)
+{
+       return dma_alloc_coherent(&dma_pdev->dev, size, handle, GFP_KERNEL);
+}
+EXPORT_SYMBOL(pasemi_dma_alloc_buf);
+
+/* pasemi_dma_free_buf - Free a buffer used for DMA
+ * @chan: Channel the buffer was allocated for
+ * @size: Size of buffer in bytes
+ * @handle: DMA handle
+ *
+ * Frees a previously allocated buffer.
+ */
+void pasemi_dma_free_buf(struct pasemi_dmachan *chan, int size,
+                        dma_addr_t *handle)
+{
+       dma_free_coherent(&dma_pdev->dev, size, handle, GFP_KERNEL);
+}
+EXPORT_SYMBOL(pasemi_dma_free_buf);
+
+static void *map_onedev(struct pci_dev *p, int index)
+{
+       struct device_node *dn;
+       void __iomem *ret;
+
+       dn = pci_device_to_OF_node(p);
+       if (!dn)
+               goto fallback;
+
+       ret = of_iomap(dn, index);
+       if (!ret)
+               goto fallback;
+
+       return ret;
+fallback:
+       /* This is hardcoded and ugly, but we have some firmware versions
+        * that don't provide the register space in the device tree. Luckily
+        * they are at well-known locations so we can just do the math here.
+        */
+       return ioremap(0xe0000000 + (p->devfn << 12), 0x2000);
+}
+
+/* pasemi_dma_init - Initialize the PA Semi DMA library
+ *
+ * This function initializes the DMA library. It must be called before
+ * any other function in the library.
+ *
+ * Returns 0 on success, errno on failure.
+ */
+int pasemi_dma_init(void)
+{
+       static spinlock_t init_lock = SPIN_LOCK_UNLOCKED;
+       struct pci_dev *iob_pdev;
+       struct pci_dev *pdev;
+       struct resource res;
+       struct device_node *dn;
+       int i, intf, err = 0;
+       u32 tmp;
+
+       if (!machine_is(pasemi))
+               return -ENODEV;
+
+       spin_lock(&init_lock);
+
+       /* Make sure we haven't already initialized */
+       if (dma_pdev)
+               goto out;
+
+       iob_pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa001, NULL);
+       if (!iob_pdev) {
+               BUG();
+               printk(KERN_WARNING "Can't find I/O Bridge\n");
+               err = -ENODEV;
+               goto out;
+       }
+       iob_regs = map_onedev(iob_pdev, 0);
+
+       dma_pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa007, NULL);
+       if (!dma_pdev) {
+               BUG();
+               printk(KERN_WARNING "Can't find DMA controller\n");
+               err = -ENODEV;
+               goto out;
+       }
+       dma_regs = map_onedev(dma_pdev, 0);
+       base_hw_irq = virq_to_hw(dma_pdev->irq);
+
+       pci_read_config_dword(dma_pdev, PAS_DMA_CAP_TXCH, &tmp);
+       num_txch = (tmp & PAS_DMA_CAP_TXCH_TCHN_M) >> PAS_DMA_CAP_TXCH_TCHN_S;
+
+       pci_read_config_dword(dma_pdev, PAS_DMA_CAP_RXCH, &tmp);
+       num_rxch = (tmp & PAS_DMA_CAP_RXCH_RCHN_M) >> PAS_DMA_CAP_RXCH_RCHN_S;
+
+       intf = 0;
+       for (pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa006, NULL);
+            pdev;
+            pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa006, pdev))
+               mac_regs[intf++] = map_onedev(pdev, 0);
+
+       pci_dev_put(pdev);
+
+       for (pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa005, NULL);
+            pdev;
+            pdev = pci_get_device(PCI_VENDOR_ID_PASEMI, 0xa005, pdev))
+               mac_regs[intf++] = map_onedev(pdev, 0);
+
+       pci_dev_put(pdev);
+
+       dn = pci_device_to_OF_node(iob_pdev);
+       if (dn)
+               err = of_address_to_resource(dn, 1, &res);
+       if (!dn || err) {
+               /* Fallback for old firmware */
+               res.start = 0xfd800000;
+               res.end = res.start + 0x1000;
+       }
+       dma_status = __ioremap(res.start, res.end-res.start, 0);
+       pci_dev_put(iob_pdev);
+
+       for (i = 0; i < MAX_TXCH; i++)
+               __set_bit(i, txch_free);
+
+       for (i = 0; i < MAX_RXCH; i++)
+               __set_bit(i, rxch_free);
+
+       printk(KERN_INFO "PA Semi PWRficient DMA library initialized "
+               "(%d tx, %d rx channels)\n", num_txch, num_rxch);
+
+out:
+       spin_unlock(&init_lock);
+       return err;
+}
+EXPORT_SYMBOL(pasemi_dma_init);
index 516acab..58b344c 100644 (file)
@@ -9,6 +9,7 @@ extern void __devinit pas_pci_dma_dev_setup(struct pci_dev *dev);
 extern void __iomem *pasemi_pci_getcfgaddr(struct pci_dev *dev, int offset);
 
 extern void __init alloc_iobmap_l2(void);
+extern void __init pasemi_map_registers(void);
 
 /* Power savings modes, implemented in asm */
 extern void idle_spin(void);
index 615b658..06bb5b7 100644 (file)
@@ -272,7 +272,7 @@ scc_enet_timeout(struct net_device *dev)
  * This is called from the CPM handler, not the MPC core interrupt.
  */
 static irqreturn_t
-scc_enet_interrupt(int irq, void * dev_id)
+scc_enet_interrupt(int irq, void *dev_id)
 {
        struct  net_device *dev = dev_id;
        volatile struct scc_enet_private *cep;
@@ -280,7 +280,7 @@ scc_enet_interrupt(int irq, void * dev_id)
        ushort  int_events;
        int     must_restart;
 
-       cep = (struct scc_enet_private *)dev->priv;
+       cep = dev->priv;
 
        /* Get the interrupt events that caused us to be here.
        */
index 6f3ed6a..a3a27da 100644 (file)
@@ -524,7 +524,7 @@ fcc_enet_timeout(struct net_device *dev)
 
 /* The interrupt handler. */
 static irqreturn_t
-fcc_enet_interrupt(int irq, void * dev_id)
+fcc_enet_interrupt(int irq, void *dev_id)
 {
        struct  net_device *dev = dev_id;
        volatile struct fcc_enet_private *cep;
@@ -532,7 +532,7 @@ fcc_enet_interrupt(int irq, void * dev_id)
        ushort  int_events;
        int     must_restart;
 
-       cep = (struct fcc_enet_private *)dev->priv;
+       cep = dev->priv;
 
        /* Get the interrupt events that caused us to be here.
        */
index b34b382..7b44a59 100644 (file)
@@ -2163,7 +2163,6 @@ static int __devinit amb_init (amb_dev * dev)
 static void setup_dev(amb_dev *dev, struct pci_dev *pci_dev) 
 {
       unsigned char pool;
-      memset (dev, 0, sizeof(amb_dev));
       
       // set up known dev items straight away
       dev->pci_dev = pci_dev; 
@@ -2253,7 +2252,7 @@ static int __devinit amb_probe(struct pci_dev *pci_dev, const struct pci_device_
                goto out_disable;
        }
 
-       dev = kmalloc (sizeof(amb_dev), GFP_KERNEL);
+       dev = kzalloc(sizeof(amb_dev), GFP_KERNEL);
        if (!dev) {
                PRINTK (KERN_ERR, "out of memory!");
                err = -ENOMEM;
index 3b64a99..2e3395b 100644 (file)
@@ -1,5 +1,3 @@
-/* $Id: he.c,v 1.18 2003/05/06 22:57:15 chas Exp $ */
-
 /*
 
   he.c
 #define HPRINTK(fmt,args...)   do { } while (0)
 #endif /* HE_DEBUG */
 
-/* version definition */
-
-static char *version = "$Id: he.c,v 1.18 2003/05/06 22:57:15 chas Exp $";
-
 /* declarations */
 
 static int he_open(struct atm_vcc *vcc);
@@ -366,7 +360,7 @@ he_init_one(struct pci_dev *pci_dev, const struct pci_device_id *pci_ent)
        struct he_dev *he_dev = NULL;
        int err = 0;
 
-       printk(KERN_INFO "he: %s\n", version);
+       printk(KERN_INFO "ATM he driver\n");
 
        if (pci_enable_device(pci_dev))
                return -EIO;
@@ -1643,6 +1637,8 @@ he_stop(struct he_dev *he_dev)
 
        if (he_dev->rbpl_base) {
 #ifdef USE_RBPL_POOL
+               int i;
+
                for (i = 0; i < CONFIG_RBPL_SIZE; ++i) {
                        void *cpuaddr = he_dev->rbpl_virt[i].virt;
                        dma_addr_t dma_handle = he_dev->rbpl_base[i].phys;
@@ -1665,6 +1661,8 @@ he_stop(struct he_dev *he_dev)
 #ifdef USE_RBPS
        if (he_dev->rbps_base) {
 #ifdef USE_RBPS_POOL
+               int i;
+
                for (i = 0; i < CONFIG_RBPS_SIZE; ++i) {
                        void *cpuaddr = he_dev->rbps_virt[i].virt;
                        dma_addr_t dma_handle = he_dev->rbps_base[i].phys;
@@ -2933,7 +2931,7 @@ he_proc_read(struct atm_dev *dev, loff_t *pos, char *page)
 
        left = *pos;
        if (!left--)
-               return sprintf(page, "%s\n", version);
+               return sprintf(page, "ATM he driver\n");
 
        if (!left--)
                return sprintf(page, "%s%s\n\n",
index 12ceed5..5732ca3 100644 (file)
@@ -104,7 +104,6 @@ int cn_queue_add_callback(struct cn_queue_dev *dev, char *name, struct cb_id *id
                return -EINVAL;
        }
 
-       cbq->nls = dev->nls;
        cbq->seq = 0;
        cbq->group = cbq->id.id.idx;
 
@@ -146,7 +145,6 @@ struct cn_queue_dev *cn_queue_alloc_dev(char *name, struct sock *nls)
        spin_lock_init(&dev->queue_lock);
 
        dev->nls = nls;
-       dev->netlink_groups = 0;
 
        dev->cn_queue = create_workqueue(dev->name);
        if (!dev->cn_queue) {
index bf9716b..fea2d3e 100644 (file)
@@ -88,6 +88,7 @@ int cn_netlink_send(struct cn_msg *msg, u32 __group, gfp_t gfp_mask)
                        if (cn_cb_equal(&__cbq->id.id, &msg->id)) {
                                found = 1;
                                group = __cbq->group;
+                               break;
                        }
                }
                spin_unlock_bh(&dev->cbdev->queue_lock);
@@ -181,33 +182,14 @@ static int cn_call_callback(struct cn_msg *msg, void (*destruct_data)(void *), v
 }
 
 /*
- * Skb receive helper - checks skb and msg size and calls callback
- * helper.
- */
-static int __cn_rx_skb(struct sk_buff *skb, struct nlmsghdr *nlh)
-{
-       u32 pid, uid, seq, group;
-       struct cn_msg *msg;
-
-       pid = NETLINK_CREDS(skb)->pid;
-       uid = NETLINK_CREDS(skb)->uid;
-       seq = nlh->nlmsg_seq;
-       group = NETLINK_CB((skb)).dst_group;
-       msg = NLMSG_DATA(nlh);
-
-       return cn_call_callback(msg, (void (*)(void *))kfree_skb, skb);
-}
-
-/*
  * Main netlink receiving function.
  *
- * It checks skb and netlink header sizes and calls the skb receive
- * helper with a shared skb.
+ * It checks skb, netlink header and msg sizes, and calls callback helper.
  */
 static void cn_rx_skb(struct sk_buff *__skb)
 {
+       struct cn_msg *msg;
        struct nlmsghdr *nlh;
-       u32 len;
        int err;
        struct sk_buff *skb;
 
@@ -223,11 +205,8 @@ static void cn_rx_skb(struct sk_buff *__skb)
                        return;
                }
 
-               len = NLMSG_ALIGN(nlh->nlmsg_len);
-               if (len > skb->len)
-                       len = skb->len;
-
-               err = __cn_rx_skb(skb, nlh);
+               msg = NLMSG_DATA(nlh);
+               err = cn_call_callback(msg, (void (*)(void *))kfree_skb, skb);
                if (err < 0)
                        kfree_skb(skb);
        }
@@ -441,8 +420,7 @@ static int __devinit cn_init(void)
 
        dev->cbdev = cn_queue_alloc_dev("cqueue", dev->nls);
        if (!dev->cbdev) {
-               if (dev->nls->sk_socket)
-                       sock_release(dev->nls->sk_socket);
+               netlink_kernel_release(dev->nls);
                return -EINVAL;
        }
        
@@ -452,8 +430,7 @@ static int __devinit cn_init(void)
        if (err) {
                cn_already_initialized = 0;
                cn_queue_free_dev(dev->cbdev);
-               if (dev->nls->sk_socket)
-                       sock_release(dev->nls->sk_socket);
+               netlink_kernel_release(dev->nls);
                return -EINVAL;
        }
 
@@ -468,8 +445,7 @@ static void __devexit cn_fini(void)
 
        cn_del_callback(&dev->id);
        cn_queue_free_dev(dev->cbdev);
-       if (dev->nls->sk_socket)
-               sock_release(dev->nls->sk_socket);
+       netlink_kernel_release(dev->nls);
 }
 
 subsys_initcall(cn_init);
index 5381c80..a58ad8a 100644 (file)
@@ -110,7 +110,7 @@ int rdma_translate_ip(struct sockaddr *addr, struct rdma_dev_addr *dev_addr)
        __be32 ip = ((struct sockaddr_in *) addr)->sin_addr.s_addr;
        int ret;
 
-       dev = ip_dev_find(ip);
+       dev = ip_dev_find(&init_net, ip);
        if (!dev)
                return -EADDRNOTAVAIL;
 
@@ -158,7 +158,7 @@ static void addr_send_arp(struct sockaddr_in *dst_in)
 
        memset(&fl, 0, sizeof fl);
        fl.nl_u.ip4_u.daddr = dst_ip;
-       if (ip_route_output_key(&rt, &fl))
+       if (ip_route_output_key(&init_net, &rt, &fl))
                return;
 
        neigh_event_send(rt->u.dst.neighbour, NULL);
@@ -179,7 +179,7 @@ static int addr_resolve_remote(struct sockaddr_in *src_in,
        memset(&fl, 0, sizeof fl);
        fl.nl_u.ip4_u.daddr = dst_ip;
        fl.nl_u.ip4_u.saddr = src_ip;
-       ret = ip_route_output_key(&rt, &fl);
+       ret = ip_route_output_key(&init_net, &rt, &fl);
        if (ret)
                goto out;
 
@@ -261,15 +261,15 @@ static int addr_resolve_local(struct sockaddr_in *src_in,
        __be32 dst_ip = dst_in->sin_addr.s_addr;
        int ret;
 
-       dev = ip_dev_find(dst_ip);
+       dev = ip_dev_find(&init_net, dst_ip);
        if (!dev)
                return -EADDRNOTAVAIL;
 
-       if (ZERONET(src_ip)) {
+       if (ipv4_is_zeronet(src_ip)) {
                src_in->sin_family = dst_in->sin_family;
                src_in->sin_addr.s_addr = dst_ip;
                ret = rdma_copy_addr(addr, dev, dev->dev_addr);
-       } else if (LOOPBACK(src_ip)) {
+       } else if (ipv4_is_loopback(src_ip)) {
                ret = rdma_translate_ip((struct sockaddr *)dst_in, addr);
                if (!ret)
                        memcpy(addr->dst_dev_addr, dev->dev_addr, MAX_ADDR_LEN);
index 637efea..1eff1b2 100644 (file)
@@ -630,7 +630,8 @@ static inline int cma_zero_addr(struct sockaddr *addr)
        struct in6_addr *ip6;
 
        if (addr->sa_family == AF_INET)
-               return ZERONET(((struct sockaddr_in *) addr)->sin_addr.s_addr);
+               return ipv4_is_zeronet(
+                       ((struct sockaddr_in *)addr)->sin_addr.s_addr);
        else {
                ip6 = &((struct sockaddr_in6 *) addr)->sin6_addr;
                return (ip6->s6_addr32[0] | ip6->s6_addr32[1] |
@@ -640,7 +641,7 @@ static inline int cma_zero_addr(struct sockaddr *addr)
 
 static inline int cma_loopback_addr(struct sockaddr *addr)
 {
-       return LOOPBACK(((struct sockaddr_in *) addr)->sin_addr.s_addr);
+       return ipv4_is_loopback(((struct sockaddr_in *) addr)->sin_addr.s_addr);
 }
 
 static inline int cma_any_addr(struct sockaddr *addr)
@@ -1288,7 +1289,7 @@ static int iw_conn_req_handler(struct iw_cm_id *cm_id,
        atomic_inc(&conn_id->dev_remove);
        conn_id->state = CMA_CONNECT;
 
-       dev = ip_dev_find(iw_event->local_addr.sin_addr.s_addr);
+       dev = ip_dev_find(&init_net, iw_event->local_addr.sin_addr.s_addr);
        if (!dev) {
                ret = -EADDRNOTAVAIL;
                cma_enable_remove(conn_id);
index f8cb0fe..e9a08fa 100644 (file)
@@ -332,7 +332,7 @@ static struct rtable *find_route(struct t3cdev *dev, __be32 local_ip,
                          }
        };
 
-       if (ip_route_output_flow(&rt, &fl, NULL, 0))
+       if (ip_route_output_flow(&init_net, &rt, &fl, NULL, 0))
                return NULL;
        return rt;
 }
index be71868..7d25368 100644 (file)
@@ -17,7 +17,7 @@
        Annapolis MD 21403
 
     Fixed (again!) the missing interrupt locking on TX/RX shifting.
-               Alan Cox <Alan.Cox@linux.org>
+       Alan Cox <Alan.Cox@linux.org>
 
     Removed calls to init_etherdev since they are no longer needed, and
     cleaned up modularization just a bit. The driver still allows only
     the board. Now getting 150K/second FTP with a 3c501 card. Still playing
     with a TX-TX optimisation to see if we can touch 180-200K/second as seems
     theoretically maximum.
-               19950402 Alan Cox <Alan.Cox@linux.org>
+               19950402 Alan Cox <Alan.Cox@linux.org>
 
     Cleaned up for 2.3.x because we broke SMP now.
-               20000208 Alan Cox <alan@redhat.com>
+               20000208 Alan Cox <alan@redhat.com>
 
     Check up pass for 2.5. Nothing significant changed
-               20021009 Alan Cox <alan@redhat.com>
+               20021009 Alan Cox <alan@redhat.com>
 
     Fixed zero fill corner case
-               20030104 Alan Cox <alan@redhat.com>
+               20030104 Alan Cox <alan@redhat.com>
 
 
    For the avoidance of doubt the "preferred form" of this code is one which
@@ -139,8 +139,8 @@ static const char version[] =
  *     The boilerplate probe code.
  */
 
-static int io=0x280;
-static int irq=5;
+static int io = 0x280;
+static int irq = 5;
 static int mem_start;
 
 /**
@@ -229,8 +229,7 @@ static int __init el1_probe1(struct net_device *dev, int ioaddr)
         *      Read the station address PROM data from the special port.
         */
 
-       for (i = 0; i < 6; i++)
-       {
+       for (i = 0; i < 6; i++) {
                outw(i, ioaddr + EL1_DATAPTR);
                station_addr[i] = inb(ioaddr + EL1_SAPROM);
        }
@@ -240,28 +239,24 @@ static int __init el1_probe1(struct net_device *dev, int ioaddr)
         */
 
        if (station_addr[0] == 0x02  &&  station_addr[1] == 0x60
-               && station_addr[2] == 0x8c)
-       {
+                                               && station_addr[2] == 0x8c)
                mname = "3c501";
-       } else if (station_addr[0] == 0x00  &&  station_addr[1] == 0x80
-       && station_addr[2] == 0xC8)
-       {
+       else if (station_addr[0] == 0x00  &&  station_addr[1] == 0x80
+                                               && station_addr[2] == 0xC8)
                mname = "NP943";
-       }
-       else {
+       else {
                release_region(ioaddr, EL1_IO_EXTENT);
                return -ENODEV;
        }
 
        /*
-        *      We auto-IRQ by shutting off the interrupt line and letting it float
-        *      high.
+        *      We auto-IRQ by shutting off the interrupt line and letting it
+        *      float high.
         */
 
        dev->irq = irq;
 
-       if (dev->irq < 2)
-       {
+       if (dev->irq < 2) {
                unsigned long irq_mask;
 
                irq_mask = probe_irq_on();
@@ -274,8 +269,7 @@ static int __init el1_probe1(struct net_device *dev, int ioaddr)
                mdelay(20);
                autoirq = probe_irq_off(irq_mask);
 
-               if (autoirq == 0)
-               {
+               if (autoirq == 0) {
                        printk(KERN_WARNING "%s probe at %#x failed to detect IRQ line.\n",
                                mname, ioaddr);
                        release_region(ioaddr, EL1_IO_EXTENT);
@@ -292,7 +286,8 @@ static int __init el1_probe1(struct net_device *dev, int ioaddr)
        if (autoirq)
                dev->irq = autoirq;
 
-       printk(KERN_INFO "%s: %s EtherLink at %#lx, using %sIRQ %d.\n", dev->name, mname, dev->base_addr,
+       printk(KERN_INFO "%s: %s EtherLink at %#lx, using %sIRQ %d.\n",
+                       dev->name, mname, dev->base_addr,
                        autoirq ? "auto":"assigned ", dev->irq);
 
 #ifdef CONFIG_IP_MULTICAST
@@ -343,7 +338,8 @@ static int el_open(struct net_device *dev)
        if (el_debug > 2)
                printk(KERN_DEBUG "%s: Doing el_open()...", dev->name);
 
-       if ((retval = request_irq(dev->irq, &el_interrupt, 0, dev->name, dev)))
+       retval = request_irq(dev->irq, &el_interrupt, 0, dev->name, dev);
+       if (retval)
                return retval;
 
        spin_lock_irqsave(&lp->lock, flags);
@@ -371,8 +367,9 @@ static void el_timeout(struct net_device *dev)
        int ioaddr = dev->base_addr;
 
        if (el_debug)
-               printk (KERN_DEBUG "%s: transmit timed out, txsr %#2x axsr=%02x rxsr=%02x.\n",
-                       dev->name, inb(TX_STATUS), inb(AX_STATUS), inb(RX_STATUS));
+               printk(KERN_DEBUG "%s: transmit timed out, txsr %#2x axsr=%02x rxsr=%02x.\n",
+                       dev->name, inb(TX_STATUS),
+                       inb(AX_STATUS), inb(RX_STATUS));
        dev->stats.tx_errors++;
        outb(TX_NORM, TX_CMD);
        outb(RX_NORM, RX_CMD);
@@ -425,8 +422,7 @@ static int el_start_xmit(struct sk_buff *skb, struct net_device *dev)
 
        netif_stop_queue(dev);
 
-       do
-       {
+       do {
                int len = skb->len;
                int pad = 0;
                int gp_start;
@@ -435,10 +431,10 @@ static int el_start_xmit(struct sk_buff *skb, struct net_device *dev)
                if (len < ETH_ZLEN)
                        pad = ETH_ZLEN - len;
 
-               gp_start = 0x800 - ( len + pad );
+               gp_start = 0x800 - (len + pad);
 
                lp->tx_pkt_start = gp_start;
-               lp->collisions = 0;
+               lp->collisions = 0;
 
                dev->stats.tx_bytes += skb->len;
 
@@ -455,37 +451,42 @@ static int el_start_xmit(struct sk_buff *skb, struct net_device *dev)
                lp->txing = 1;
 
                /*
-                *      Turn interrupts back on while we spend a pleasant afternoon
-                *      loading bytes into the board
+                *      Turn interrupts back on while we spend a pleasant
+                *      afternoon loading bytes into the board
                 */
 
                spin_unlock_irqrestore(&lp->lock, flags);
 
-               outw(0x00, RX_BUF_CLR);         /* Set rx packet area to 0. */
-               outw(gp_start, GP_LOW);         /* aim - packet will be loaded into buffer start */
-               outsb(DATAPORT,buf,len);        /* load buffer (usual thing each byte increments the pointer) */
+               /* Set rx packet area to 0. */
+               outw(0x00, RX_BUF_CLR);
+               /* aim - packet will be loaded into buffer start */
+               outw(gp_start, GP_LOW);
+               /* load buffer (usual thing each byte increments the pointer) */
+               outsb(DATAPORT, buf, len);
                if (pad) {
-                       while(pad--)            /* Zero fill buffer tail */
+                       while (pad--)           /* Zero fill buffer tail */
                                outb(0, DATAPORT);
                }
-               outw(gp_start, GP_LOW);         /* the board reuses the same register */
+               /* the board reuses the same register */
+               outw(gp_start, GP_LOW);
 
-               if(lp->loading != 2)
-               {
-                       outb(AX_XMIT, AX_CMD);          /* fire ... Trigger xmit.  */
-                       lp->loading=0;
+               if (lp->loading != 2) {
+                       /* fire ... Trigger xmit.  */
+                       outb(AX_XMIT, AX_CMD);
+                       lp->loading = 0;
                        dev->trans_start = jiffies;
                        if (el_debug > 2)
                                printk(KERN_DEBUG " queued xmit.\n");
-                       dev_kfree_skb (skb);
+                       dev_kfree_skb(skb);
                        return 0;
                }
                /* A receive upset our load, despite our best efforts */
-               if(el_debug>2)
-                       printk(KERN_DEBUG "%s: burped during tx load.\n", dev->name);
+               if (el_debug > 2)
+                       printk(KERN_DEBUG "%s: burped during tx load.\n",
+                               dev->name);
                spin_lock_irqsave(&lp->lock, flags);
        }
-       while(1);
+       while (1);
 
 }
 
@@ -534,64 +535,59 @@ static irqreturn_t el_interrupt(int irq, void *dev_id)
         */
 
        if (el_debug > 3)
-               printk(KERN_DEBUG "%s: el_interrupt() aux=%#02x", dev->name, axsr);
-
-        if(lp->loading==1 && !lp->txing)
-               printk(KERN_WARNING "%s: Inconsistent state loading while not in tx\n",
-                       dev->name);
-
-       if (lp->txing)
-       {
+               printk(KERN_DEBUG "%s: el_interrupt() aux=%#02x",
+                                                       dev->name, axsr);
 
-               /*
-                *      Board in transmit mode. May be loading. If we are
-                *      loading we shouldn't have got this.
-                */
+       if (lp->loading == 1 && !lp->txing)
+               printk(KERN_WARNING "%s: Inconsistent state loading while not in tx\n",
+                       dev->name);
 
+       if (lp->txing) {
+               /*
+                *      Board in transmit mode. May be loading. If we are
+                *      loading we shouldn't have got this.
+                */
                int txsr = inb(TX_STATUS);
 
-               if(lp->loading==1)
-               {
-                       if(el_debug > 2)
-                       {
-                               printk(KERN_DEBUG "%s: Interrupt while loading [", dev->name);
-                               printk(KERN_DEBUG " txsr=%02x gp=%04x rp=%04x]\n", txsr, inw(GP_LOW),inw(RX_LOW));
+               if (lp->loading == 1) {
+                       if (el_debug > 2) {
+                               printk(KERN_DEBUG "%s: Interrupt while loading [",
+                                       dev->name);
+                               printk(" txsr=%02x gp=%04x rp=%04x]\n",
+                                       txsr, inw(GP_LOW), inw(RX_LOW));
                        }
-                       lp->loading=2;          /* Force a reload */
+                       /* Force a reload */
+                       lp->loading = 2;
                        spin_unlock(&lp->lock);
                        goto out;
                }
-
                if (el_debug > 6)
-                       printk(KERN_DEBUG " txsr=%02x gp=%04x rp=%04x", txsr, inw(GP_LOW),inw(RX_LOW));
+                       printk(KERN_DEBUG " txsr=%02x gp=%04x rp=%04x",
+                                       txsr, inw(GP_LOW), inw(RX_LOW));
 
-               if ((axsr & 0x80) && (txsr & TX_READY) == 0)
-               {
+               if ((axsr & 0x80) && (txsr & TX_READY) == 0) {
                        /*
-                        *      FIXME: is there a logic to whether to keep on trying or
-                        *      reset immediately ?
+                        *      FIXME: is there a logic to whether to keep
+                        *      on trying or reset immediately ?
                         */
-                       if(el_debug>1)
-                               printk(KERN_DEBUG "%s: Unusual interrupt during Tx, txsr=%02x axsr=%02x"
-                                       " gp=%03x rp=%03x.\n", dev->name, txsr, axsr,
-                       inw(ioaddr + EL1_DATAPTR), inw(ioaddr + EL1_RXPTR));
+                       if (el_debug > 1)
+                               printk(KERN_DEBUG "%s: Unusual interrupt during Tx, txsr=%02x axsr=%02x gp=%03x rp=%03x.\n",
+                                       dev->name, txsr, axsr,
+                                       inw(ioaddr + EL1_DATAPTR),
+                                       inw(ioaddr + EL1_RXPTR));
                        lp->txing = 0;
                        netif_wake_queue(dev);
-               }
-               else if (txsr & TX_16COLLISIONS)
-               {
+               } else if (txsr & TX_16COLLISIONS) {
                        /*
                         *      Timed out
                         */
                        if (el_debug)
-                               printk (KERN_DEBUG "%s: Transmit failed 16 times, Ethernet jammed?\n",dev->name);
+                               printk(KERN_DEBUG "%s: Transmit failed 16 times, Ethernet jammed?\n", dev->name);
                        outb(AX_SYS, AX_CMD);
                        lp->txing = 0;
                        dev->stats.tx_aborted_errors++;
                        netif_wake_queue(dev);
-               }
-               else if (txsr & TX_COLLISION)
-               {
+               } else if (txsr & TX_COLLISION) {
                        /*
                         *      Retrigger xmit.
                         */
@@ -599,7 +595,8 @@ static irqreturn_t el_interrupt(int irq, void *dev_id)
                        if (el_debug > 6)
                                printk(KERN_DEBUG " retransmitting after a collision.\n");
                        /*
-                        *      Poor little chip can't reset its own start pointer
+                        *      Poor little chip can't reset its own start
+                        *      pointer
                         */
 
                        outb(AX_SYS, AX_CMD);
@@ -608,53 +605,45 @@ static irqreturn_t el_interrupt(int irq, void *dev_id)
                        dev->stats.collisions++;
                        spin_unlock(&lp->lock);
                        goto out;
-               }
-               else
-               {
+               } else {
                        /*
                         *      It worked.. we will now fall through and receive
                         */
                        dev->stats.tx_packets++;
                        if (el_debug > 6)
                                printk(KERN_DEBUG " Tx succeeded %s\n",
-                                       (txsr & TX_RDY) ? "." : "but tx is busy!");
+                                       (txsr & TX_RDY) ? "." : "but tx is busy!");
                        /*
                         *      This is safe the interrupt is atomic WRT itself.
                         */
-
                        lp->txing = 0;
-                       netif_wake_queue(dev);  /* In case more to transmit */
+                       /* In case more to transmit */
+                       netif_wake_queue(dev);
                }
-       }
-       else
-       {
-               /*
-                *      In receive mode.
-                */
+       } else {
+               /*
+                *      In receive mode.
+                */
 
                int rxsr = inb(RX_STATUS);
                if (el_debug > 5)
-                       printk(KERN_DEBUG " rxsr=%02x txsr=%02x rp=%04x", rxsr, inb(TX_STATUS),inw(RX_LOW));
+                       printk(KERN_DEBUG " rxsr=%02x txsr=%02x rp=%04x", rxsr, inb(TX_STATUS), inw(RX_LOW));
                /*
                 *      Just reading rx_status fixes most errors.
                 */
                if (rxsr & RX_MISSED)
                        dev->stats.rx_missed_errors++;
-               else if (rxsr & RX_RUNT)
-               {       /* Handled to avoid board lock-up. */
+               else if (rxsr & RX_RUNT) {
+                       /* Handled to avoid board lock-up. */
                        dev->stats.rx_length_errors++;
                        if (el_debug > 5)
                                printk(KERN_DEBUG " runt.\n");
-               }
-               else if (rxsr & RX_GOOD)
-               {
+               } else if (rxsr & RX_GOOD) {
                        /*
                         *      Receive worked.
                         */
                        el_receive(dev);
-               }
-               else
-               {
+               } else {
                        /*
                         *      Nothing?  Something is broken!
                         */
@@ -702,8 +691,7 @@ static void el_receive(struct net_device *dev)
        if (el_debug > 4)
                printk(KERN_DEBUG " el_receive %d.\n", pkt_len);
 
-       if ((pkt_len < 60)  ||  (pkt_len > 1536))
-       {
+       if (pkt_len < 60 || pkt_len > 1536) {
                if (el_debug)
                        printk(KERN_DEBUG "%s: bogus packet, length=%d\n", dev->name, pkt_len);
                dev->stats.rx_over_errors++;
@@ -722,26 +710,23 @@ static void el_receive(struct net_device *dev)
         */
 
        outw(0x00, GP_LOW);
-       if (skb == NULL)
-       {
+       if (skb == NULL) {
                printk(KERN_INFO "%s: Memory squeeze, dropping packet.\n", dev->name);
                dev->stats.rx_dropped++;
                return;
-       }
-       else
-       {
-               skb_reserve(skb,2);     /* Force 16 byte alignment */
+       } else {
+               skb_reserve(skb, 2);    /* Force 16 byte alignment */
                /*
                 *      The read increments through the bytes. The interrupt
                 *      handler will fix the pointer when it returns to
                 *      receive mode.
                 */
-               insb(DATAPORT, skb_put(skb,pkt_len), pkt_len);
-               skb->protocol=eth_type_trans(skb,dev);
+               insb(DATAPORT, skb_put(skb, pkt_len), pkt_len);
+               skb->protocol = eth_type_trans(skb, dev);
                netif_rx(skb);
                dev->last_rx = jiffies;
                dev->stats.rx_packets++;
-   &