summaryrefslogtreecommitdiffstats
path: root/net
Commit message (Collapse)AuthorAgeFilesLines
* netns: embed ip6_dst_ops directlyAlexey Dobriyan2009-09-011-21/+13
| | | | | | | | | | | | | | | | | struct net::ipv6.ip6_dst_ops is separatedly dynamically allocated, but there is no fundamental reason for it. Embed it directly into struct netns_ipv6. For that: * move struct dst_ops into separate header to fix circular dependencies I honestly tried not to, it's pretty impossible to do other way * drop dynamical allocation, allocate together with netns For a change, remove struct dst_ops::dst_net, it's deducible by using container_of() given dst_ops pointer. Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* Revert Backoff [v3]: Calculate TCP's connection close threshold as a time value.Damian Lukowski2009-09-011-4/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | RFC 1122 specifies two threshold values R1 and R2 for connection timeouts, which may represent a number of allowed retransmissions or a timeout value. Currently linux uses sysctl_tcp_retries{1,2} to specify the thresholds in number of allowed retransmissions. For any desired threshold R2 (by means of time) one can specify tcp_retries2 (by means of number of retransmissions) such that TCP will not time out earlier than R2. This is the case, because the RTO schedule follows a fixed pattern, namely exponential backoff. However, the RTO behaviour is not predictable any more if RTO backoffs can be reverted, as it is the case in the draft "Make TCP more Robust to Long Connectivity Disruptions" (http://tools.ietf.org/html/draft-zimmermann-tcp-lcd). In the worst case TCP would time out a connection after 3.2 seconds, if the initial RTO equaled MIN_RTO and each backoff has been reverted. This patch introduces a function retransmits_timed_out(N), which calculates the timeout of a TCP connection, assuming an initial RTO of MIN_RTO and N unsuccessful, exponentially backed-off retransmissions. Whenever timeout decisions are made by comparing the retransmission counter to some value N, this function can be used, instead. The meaning of tcp_retries2 will be changed, as many more RTO retransmissions can occur than the value indicates. However, it yields a timeout which is similar to the one of an unpatched, exponentially backing off TCP in the same scenario. As no application could rely on an RTO greater than MIN_RTO, there should be no risk of a regression. Signed-off-by: Damian Lukowski <damian@tvk.rwth-aachen.de> Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: David S. Miller <davem@davemloft.net>
* Revert Backoff [v3]: Revert RTO on ICMP destination unreachableDamian Lukowski2009-09-013-4/+40
| | | | | | | | | | | | | | | | | | | | | | | | | | Here, an ICMP host/network unreachable message, whose payload fits to TCP's SND.UNA, is taken as an indication that the RTO retransmission has not been lost due to congestion, but because of a route failure somewhere along the path. With true congestion, a router won't trigger such a message and the patched TCP will operate as standard TCP. This patch reverts one RTO backoff, if an ICMP host/network unreachable message, whose payload fits to TCP's SND.UNA, arrives. Based on the new RTO, the retransmission timer is reset to reflect the remaining time, or - if the revert clocked out the timer - a retransmission is sent out immediately. Backoffs are only reverted, if TCP is in RTO loss recovery, i.e. if there have been retransmissions and reversible backoffs, already. Changes from v2: 1) Renaming of skb in tcp_v4_err() moved to another patch. 2) Reintroduced tcp_bound_rto() and __tcp_set_rto(). 3) Fixed code comments. Signed-off-by: Damian Lukowski <damian@tvk.rwth-aachen.de> Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: David S. Miller <davem@davemloft.net>
* Revert Backoff [v3]: Rename skb to icmp_skb in tcp_v4_err()Damian Lukowski2009-09-011-8/+8
| | | | | | | | | | This supplementary patch renames skb to icmp_skb in tcp_v4_err() in order to disambiguate from another sk_buff variable, which will be introduced in a separate patch. Signed-off-by: Damian Lukowski <damian@tvk.rwth-aachen.de> Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: David S. Miller <davem@davemloft.net>
* dcbnl: Add implementations of dcbnl setapp/getapp commandsYi Zou2009-09-011-0/+122
| | | | | | | | | | | Implements the dcbnl netlink setapp/getapp pair. When a setapp/getapp is received, dcbnl would just pass on to dcbnl_rtnl_op.setapp/getapp that are supposed to be implemented by the low level drivers. Signed-off-by: Yi Zou <yi.zou@intel.com> Acked-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* dcbnl: Add netlink attributes for setapp/getapp to dcbnlYi Zou2009-09-011-0/+8
| | | | | | | | | | Add defines for dcbnl netlink attributes to support netlink message passing of setapp/getapp in dcbnl. Signed-off-by: Yi Zou <yi.zou@intel.com> Acked-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* vlan: Add support for net_devices_ops.ndo_fcoe_enable/_disable to VLANYi Zou2009-09-011-0/+26
| | | | | | | | | | This adds implementation of the net_devices_ops.ndo_fcoe_enable/_disable to the VLAN driver. It checks if the real_dev has support for ndo_fcoe_enable/ ndo_fcoe_disable and if so, passes on to call the associated real_dev. Signed-off-by: Yi Zou <yi.zou@intel.com> Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* wireless: convert drivers to netdev_tx_tStephen Hemminger2009-09-012-6/+8
| | | | | | | | | | | Mostly just simple conversions: * ray_cs had bogus return of NET_TX_LOCKED but driver was not using NETIF_F_LLTX * hostap and ipw2x00 had some code that returned value from a called function that also had to change to return netdev_tx_t Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* netdev: convert pseudo drivers to netdev_tx_tStephen Hemminger2009-09-011-1/+1
| | | | | | | These are all drivers that don't touch real hardware. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* irda: convert to netdev_tx_tStephen Hemminger2009-09-011-2/+4
| | | | | Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* convert hamradio drivers to netdev_txreturnt_tStephen Hemminger2009-09-012-2/+2
| | | | | Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* convert ATM drivers to netdev_tx_tStephen Hemminger2009-09-014-6/+12
| | | | | Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* netdev: convert pseudo-devices to netdev_tx_tStephen Hemminger2009-09-0116-21/+26
| | | | | Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* pkt_sched: Fix resource limiting in pfifo_fastKrishna Kumar2009-08-301-4/+4
| | | | | | | | | | | | | | | | | | pfifo_fast_enqueue has this check: if (skb_queue_len(list) < qdisc_dev(qdisc)->tx_queue_len) { which allows each band to enqueue upto tx_queue_len skbs for a total of 3*tx_queue_len skbs. I am not sure if this was the intention of limiting in qdisc. Patch compiled and 32 simultaneous netperf testing ran fine. Also: # tc -s qdisc show dev eth2 qdisc pfifo_fast 0: root bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 Sent 16835026752 bytes 373116 pkt (dropped 0, overlimits 0 requeues 25) rate 0bit 0pps backlog 0b 0p requeues 25 Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* net: convert remaining non-symbolic return values in dev_queue_xmitKrishna Kumar2009-08-301-1/+1
| | | | | | | Patch compiled and 32 simultaneous netperf testing ran fine. Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* can: use correct NET_RX_ return valuesOliver Hartkopp2009-08-301-2/+2
| | | | | | | | Dropped skb's should be documented by an appropriate return value. Use the correct NET_RX_DROP and NET_RX_SUCCESS values for that reason. Signed-off-by: Oliver Hartkopp <oliver@hartkopp.net> Signed-off-by: David S. Miller <davem@davemloft.net>
* Merge branch 'master' of ↵David S. Miller2009-08-309-146/+1419
|\ | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/holtmann/bluetooth-next-2.6
| * Bluetooth: Add support for L2CAP 'Send RRorRNR' actionGustavo F. Padovan2009-08-261-6/+14
| | | | | | | | | | | | | | | | When called, 'Send RRorRNR' should send a RNR frame if local device is busy or a RR frame otherwise. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Handle L2CAP case when the remote receiver is busyGustavo F. Padovan2009-08-261-2/+23
| | | | | | | | | | | | | | Implement all issues related to RemoteBusy in the RECV state table. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Acknowledge L2CAP packets when receiving RR-frames (F-bit=1)Gustavo F. Padovan2009-08-261-1/+5
| | | | | | | | | | | | | | | | Implement the Recv ReqSeqAndFBit event when a RR frame with F bit set is received. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add missing selection of CONFIG_CRC16 for L2CAP layerRandy Dunlap2009-08-241-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Fix net/bluetooth/l2cap.c build errors: l2cap.c:(.text+0x126035): undefined reference to `crc16' l2cap.c:(.text+0x126323): undefined reference to `crc16' l2cap.c:(.text+0x12668e): undefined reference to `crc16' l2cap.c:(.text+0x12683b): undefined reference to `crc16' l2cap.c:(.text+0x126956): undefined reference to `crc16' net/built-in.o:l2cap.c:(.text+0x129041): more undefined references to `crc16' follow Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Use proper *_unaligned_le{16,32} helpers for L2CAPGustavo F. Padovan2009-08-241-2/+2
| | | | | | | | | | | | | | Simplify more conversions to the right endian with the proper helpers. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add locking scheme to L2CAP timeout callbacksGustavo F. Padovan2009-08-241-0/+4
| | | | | | | | | | | | | | | | Avoid race conditions when accessing the L2CAP socket from within the timeout handlers. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Fix rejected connection not disconnecting ACL linkLuiz Augusto von Dentz2009-08-221-0/+41
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When using DEFER_SETUP on a RFCOMM socket, a SABM frame triggers authorization which when rejected send a DM response. This is fine according to the RFCOMM spec: the responding implementation may replace the "proper" response on the Multiplexer Control channel with a DM frame, sent on the referenced DLCI to indicate that the DLCI is not open, and that the responder would not grant a request to open it later either. But some stacks doesn't seems to cope with this leaving DLCI 0 open after receiving DM frame. To fix it properly a timer was introduced to rfcomm_session which is used to set a timeout when the last active DLC of a session is unlinked, this will give the remote stack some time to reply with a proper DISC frame on DLCI 0 avoiding both sides sending DISC to each other on stacks that follow the specification and taking care of those who don't by taking down DLCI 0. Signed-off-by: Luiz Augusto von Dentz <luiz.dentz@openbossa.org> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Full support for receiving L2CAP SREJ framesGustavo F. Padovan2009-08-221-1/+29
| | | | | | | | | | | | | | Support for receiving of SREJ frames as specified by the state table. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add support for L2CAP SREJ exceptionGustavo F. Padovan2009-08-221-24/+196
| | | | | | | | | | | | | | | | | | | | When L2CAP loses an I-frame we send a SREJ frame to the transmitter side requesting the lost packet. This patch implement all Recv I-frame events on SREJ_SENT state table except the ones that deal with SendRej (the REJ exception at receiver side is yet not implemented). Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add support for FCS option to L2CAPGustavo F. Padovan2009-08-221-6/+95
| | | | | | | | | | | | | | | | | | | | | | | | | | Implement CRC16 check for L2CAP packets. FCS is used by Streaming Mode and Enhanced Retransmission Mode and is a extra check for the packet content. Using CRC16 is the default, L2CAP won't use FCS only when both side send a "No FCS" request. Initially based on a patch from Nathan Holstein <nathan@lampreynetworks.com> Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Enable Streaming Mode for L2CAPGustavo F. Padovan2009-08-221-9/+73
| | | | | | | | | | | | | | | | Streaming Mode is helpful for the Bluetooth streaming based profiles, such as A2DP. It doesn't have any error control or flow control. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add support for Retransmission and Monitor TimersGustavo F. Padovan2009-08-221-4/+82
| | | | | | | | | | | | | | | | | | | | | | L2CAP uses retransmission and monitor timers to inquiry the other side about unacked I-frames. After sending each I-frame we (re)start the retransmission timer. If it expires, we start a monitor timer that send a S-frame with P bit set and wait for S-frame with F bit set. If monitor timer expires, try again, at a maximum of L2CAP_DEFAULT_MAX_TX. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Initial support for retransmission of packets with REJ framesGustavo F. Padovan2009-08-221-14/+43
| | | | | | | | | | | | | | | | | | | | | | | | When receiving an I-frame with unexpected txSeq, receiver side start the recovery procedure by sending a REJ S-frame to the transmitter side. So the transmitter can re-send the lost I-frame. This patch just adds a basic support for retransmission, it doesn't mean that ERTM now has full support for packet retransmission. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add support for Segmentation and Reassembly of SDUsGustavo F. Padovan2009-08-221-16/+154
| | | | | | | | | | | | | | | | | | | | | | | | | | | | ERTM should use Segmentation and Reassembly to break down a SDU in many PDUs on sending data to the other side. On sending packets we queue all 'segments' until end of segmentation and just the add them to the queue for sending. On receiving we create a new SKB with the SDU reassembled. Initially based on a patch from Nathan Holstein <nathan@lampreynetworks.com> Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add initial support for ERTM packets transfersGustavo F. Padovan2009-08-221-49/+335
| | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch adds support for ERTM transfers, without retransmission, with txWindow up to 63 and with acknowledgement of packets received. Now the packets are queued before call l2cap_do_send(), so packets couldn't be sent at the time we call l2cap_sock_sendmsg(). They will be sent in an asynchronous way on later calls of l2cap_ertm_send(). Besides if an error occurs on calling l2cap_do_send() we disconnect the channel. Initially based on a patch from Nathan Holstein <nathan@lampreynetworks.com> Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Create separate l2cap_send_disconn_req() functionGustavo F. Padovan2009-08-221-30/+16
| | | | | | | | | | | | | | | | The code for sending a disconnect request was repeated several times within L2CAP source code. So move this into its own function. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add configuration support for ERTM and Streaming modeGustavo F. Padovan2009-08-221-28/+234
| | | | | | | | | | | | | | | | | | | | | | Add support to config_req and config_rsp to configure ERTM and Streaming mode. If the remote device specifies ERTM or Streaming mode, then the same mode is proposed. Otherwise ERTM or Basic mode is used. And in case of a state 2 device, the remote device should propose the same mode. If not, then the channel gets disconnected. Signed-off-by: Gustavo F. Padovan <gustavo@las.ic.unicamp.br> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add L2CAP RFC option if ERTM is enabledMarcel Holtmann2009-08-221-3/+20
| | | | | | | | | | | | | | When trying to establish a connection with Enhanced Retransmission mode enabled, the RFC option needs to be added to the configuration. Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Allow setting of L2CAP ERTM via socket optionMarcel Holtmann2009-08-221-4/+33
| | | | | | | | | | | | | | | | | | To enable Enhanced Retransmission mode it needs to be set via a socket option. A different mode can be set on a socket, but on listen() and connect() the mode is checked and ERTM is only allowed if it is enabled via the module parameter. Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add module option to enable L2CAP ERTM supportMarcel Holtmann2009-08-221-2/+10
| | | | | | | | | | | | | | | | Since the Enhanced Retransmission mode for L2CAP is still under heavy development disable it by default and provide a module option to enable it manually for testing. Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Coding style cleanup from previous rfcomm_init bug fixMarcel Holtmann2009-08-221-13/+15
| | | | | | | | | | | | | | | | The rfcomm_init bug fix went into the kernel premature before it got fully reviewed and acknowledged by the Bluetooth maintainer. So fix up the coding style now. Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Convert hdev->req_lock to a mutexThomas Gleixner2009-08-221-1/+1
| | | | | | | | | | | | | | hdev->req_lock is used as mutex so make it a mutex. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Let HIDP grab the device reference for connectionsMarcel Holtmann2009-08-222-21/+43
| | | | | | | | | | | | | | | | | | | | The core exports the hci_conn_hold_device() and hci_conn_put_device() functions for device reference of connections. Use this to ensure that the uevents from the parent are send after the child ones. Based on a report by Brian Rogers <brian@xyzw.org> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add extra device reference counting for connectionsMarcel Holtmann2009-08-222-1/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The device model itself has no real usable reference counting at the moment and this causes problems if parents are deleted before their children. The device model itself handles the memory details of this correctly, but the uevent order is not consistent. This causes various problems for systems like HAL or even X. So until device_put() does a proper cleanup, the device for Bluetooth connection will be protected with an extra reference counting to ensure the correct order of uevents when connections are terminated. This is not an automatic feature. Higher Bluetooth layers like HIDP or BNEP should grab this new reference to ensure that their uevents are send before the ones from the parent device. Based on a report by Brian Rogers <brian@xyzw.org> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Disconnect HIDRAW devices on disconnectMarcel Holtmann2009-08-221-0/+3
| | | | | | | | | | | | | | | | | | | | | | Currently the HID subsystem will create HIDRAW devices for the transport driver, but it will not disconnect them. Until the HID subsytem gets fixed, ensure that HIDRAW and HIDDEV devices are disconnected when the Bluetooth HID device gets removed. Based on a patch from Brian Rogers <brian@xyzw.org> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Fix missing scheduling when VIRTUAL_CABLE_UNPLUG is receivedVikram Kandukuri2009-08-221-0/+1
| | | | | | | | | | | | | | | | | | | | | | There is a test case in PTS tool; PTS will send the VIRTUAL_CABLE_UNPLUG command to IUT. Then IUT should disconnect the channel and kill the HID session when it receives the command. The VIRTUAL_CABLE_UNPLUG command is parsed by HID transport, but it is not scheduled to do so. Add a call to hidp_schedule() to kill the session. Signed-off-by: Jothikumar Mothilal <jothikumar.mothilal@atheros.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
| * Bluetooth: Add proper shutdown support to SCO socketsMarcel Holtmann2009-08-221-15/+34
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The SCO sockets for Bluetooth audio setup and streaming are missing the shutdown implementation. This hasn't been a problem so far, but with a more deeper integration with PulseAudio it is important to shutdown SCO sockets properly. Also the Headset profile 1.2 has more detailed qualification tests that require that SCO and RFCOMM channels are terminated in the right order. A proper shutdown function is necessary for this. Based on a report by Johan Hedberg <johan.hedberg@nokia.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org> Tested-by: Johan Hedberg <johan.hedberg@nokia.com>
* | tipc: fix test of bearer_priority range in tipc_register_media()roel kluin2009-08-291-1/+1
| | | | | | | | | | | | | | | | For the bearer_priority to be less than TIPC_MIN_LINK_PRI and greater than TIPC_MAX_LINK_PRI is logically impossible. Signed-off-by: Roel Kluin <roel.kluin@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | can: switch to seq_fileAlexey Dobriyan2009-08-292-199/+167
| | | | | | | | | | | | | | create_proc_read_entry() is going to be removed soon. Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tcp: Remove redundant copy of MD5 authentication keyJohn Dykstra2009-08-291-23/+0
| | | | | | | | | | | | | | | | | | Remove the copy of the MD5 authentication key from tcp_check_req(). This key has already been copied by tcp_v4_syn_recv_sock() or tcp_v6_syn_recv_sock(). Signed-off-by: John Dykstra <john.dykstra1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | Speed-up pfifo_fast lookup using a private bitmapKrishna Kumar2009-08-291-22/+48
| | | | | | | | | | | | | | | | | | | | | | Maintain a per-qdisc bitmap for pfifo_fast giving availability of skbs for each band. This allows faster lookup for a skb when there are no high priority skbs. Also, it helps in (rare) cases when there are no skbs on the list, where an immediate lookup is faster than iterating through the three bands. Signed-off-by: Krishna Kumar <krkumar2@in.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
* | ipv6: Update Neighbor Cache when IPv6 RA is received on a routerDavid Ward2009-08-291-6/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When processing a received IPv6 Router Advertisement, the kernel creates or updates an IPv6 Neighbor Cache entry for the sender -- but presently this does not occur if IPv6 forwarding is enabled (net.ipv6.conf.*.forwarding = 1), or if IPv6 Router Advertisements are not accepted (net.ipv6.conf.*.accept_ra = 0), because in these cases processing of the Router Advertisement has already halted. This patch allows the Neighbor Cache to be updated in these cases, while still avoiding any modification to routes or link parameters. This continues to satisfy RFC 4861, since any entry created in the Neighbor Cache as the result of a received Router Advertisement is still placed in the STALE state. Signed-off-by: David Ward <david.ward@ll.mit.edu> Signed-off-by: David S. Miller <davem@davemloft.net>
* | tcp: fix premature termination of FIN_WAIT2 time-wait socketsOctavian Purdila2009-08-291-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There is a race condition in the time-wait sockets code that can lead to premature termination of FIN_WAIT2 and, subsequently, to RST generation when the FIN,ACK from the peer finally arrives: Time TCP header 0.000000 30755 > http [SYN] Seq=0 Win=2920 Len=0 MSS=1460 TSV=282912 TSER=0 0.000008 http > 30755 aSYN, ACK] Seq=0 Ack=1 Win=2896 Len=0 MSS=1460 TSV=... 0.136899 HEAD /1b.html?n1Lg=v1 HTTP/1.0 [Packet size limited during capture] 0.136934 HTTP/1.0 200 OK [Packet size limited during capture] 0.136945 http > 30755 [FIN, ACK] Seq=187 Ack=207 Win=2690 Len=0 TSV=270521... 0.136974 30755 > http [ACK] Seq=207 Ack=187 Win=2734 Len=0 TSV=283049 TSER=... 0.177983 30755 > http [ACK] Seq=207 Ack=188 Win=2733 Len=0 TSV=283089 TSER=... 0.238618 30755 > http [FIN, ACK] Seq=207 Ack=188 Win=2733 Len=0 TSV=283151... 0.238625 http > 30755 [RST] Seq=188 Win=0 Len=0 Say twdr->slot = 1 and we are running inet_twdr_hangman and in this instance inet_twdr_do_twkill_work returns 1. At that point we will mark slot 1 and schedule inet_twdr_twkill_work. We will also make twdr->slot = 2. Next, a connection is closed and tcp_time_wait(TCP_FIN_WAIT2, timeo) is called which will create a new FIN_WAIT2 time-wait socket and will place it in the last to be reached slot, i.e. twdr->slot = 1. At this point say inet_twdr_twkill_work will run which will start destroying the time-wait sockets in slot 1, including the just added TCP_FIN_WAIT2 one. To avoid this issue we increment the slot only if all entries in the slot have been purged. This change may delay the slots cleanup by a time-wait death row period but only if the worker thread didn't had the time to run/purge the current slot in the next period (6 seconds with default sysctl settings). However, on such a busy system even without this change we would probably see delays... Signed-off-by: Octavian Purdila <opurdila@ixiacom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
OpenPOWER on IntegriCloud