- 11 4月, 2011 1 次提交
-
-
由 Michael Smith 提交于
This makes sk_buff available for other use in fib_validate_source(). Signed-off-by: NMichael Smith <msmith@cbnco.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 08 4月, 2011 1 次提交
-
-
由 OGAWA Hirofumi 提交于
Commit 1018b5c0 ("Set rt->rt_iif more sanely on output routes.") breaks rt_is_{output,input}_route. This became the cause to return "IP_PKTINFO's ->ipi_ifindex == 0". To fix it, this does: 1) Add "int rt_route_iif;" to struct rtable 2) For input routes, always set rt_route_iif to same value as rt_iif 3) For output routes, always set rt_route_iif to zero. Set rt_iif as it is done currently. 4) Change rt_is_{output,input}_route() to test rt_route_iif Signed-off-by: NOGAWA Hirofumi <hirofumi@mail.parknet.co.jp> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 07 4月, 2011 3 次提交
-
-
由 Peter Korsgaard 提交于
The mv88e6085 is identical to the mv88e6095, except that all ports are 10/100 Mb/s, so use the existing setup code except for the cpu/dsa speed selection in _setup_port(). Signed-off-by: NPeter Korsgaard <jacmet@sunsite.dk> Acked-by: NLennert Buytenhek <buytenh@wantstofly.org> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Neil Horman 提交于
properly record sk_rxhash in ipv6 sockets (v2) Noticed while working on another project that flows to sockets which I had open on a test systems weren't getting steered properly when I had RFS enabled. Looking more closely I found that: 1) The affected sockets were all ipv6 2) They weren't getting steered because sk->sk_rxhash was never set from the incomming skbs on that socket. This was occuring because there are several points in the IPv4 tcp and udp code which save the rxhash value when a new connection is established. Those calls to sock_rps_save_rxhash were never added to the corresponding ipv6 code paths. This patch adds those calls. Tested by myself to properly enable RFS functionalty on ipv6. Change notes: v2: Filtered UDP to only arm RFS on bound sockets (Eric Dumazet) Signed-off-by: NNeil Horman <nhorman@tuxdriver.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Oliver Hartkopp 提交于
This patch removes spin_locks at CAN socket creation time by using RCU. Inspired by the discussion with Kurt van Dijck and Eric Dumazet the RCU code was partly derived from af_phonet.c Signed-off-by: NOliver Hartkopp <socketcan@hartkopp.net> Reviewed-by: NEric Dumazet <eric.dumazet@gmail.com> Acked-by: NKurt Van Dijck <kurt.van.dijck@eia.be> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 05 4月, 2011 12 次提交
-
-
由 Ben Hutchings 提交于
The ethtool ETHTOOL_PHYS_ID command runs for an arbitrarily long period of time, holding the RTNL lock. This blocks routing updates, device enumeration, and various important operations that one might want to keep running while hunting for the flashing LED. We need to drop the RTNL lock during this operation, but currently the core implementation is a thin wrapper around a driver operation and drivers may well depend upon holding the lock. Define a new driver operation 'set_phys_id' with an argument that sets the ID indicator on/off/inactive/active (the last optional, for any driver or firmware that prefers to handle blinking asynchronously). When this is defined, the ethtool core drops the lock while waiting and only acquires it around calls to this operation. Deprecate the 'phys_id' operation in favour of this. It can be removed once all in-tree drivers are converted. Signed-off-by: NBen Hutchings <bhutchings@solarflare.com>
-
由 Tom Herbert 提交于
This patch uses __copy_from_user_nocache on transmit to bypass data cache for a performance improvement. skb_add_data_nocache and skb_copy_to_page_nocache can be called by sendmsg functions to use this feature, initial support is in tcp_sendmsg. This functionality is configurable per device using ethtool. Presumably, this feature would only be useful when the driver does not touch the data. The feature is turned on by default if a device indicates that it does some form of checksum offload; it is off by default for devices that do no checksum offload or indicate no checksum is necessary. For the former case copy-checksum is probably done anyway, in the latter case the device is likely loopback in which case the no cache copy is probably not beneficial. This patch was tested using 200 instances of netperf TCP_RR with 1400 byte request and one byte reply. Platform is 16 core AMD x86. No-cache copy disabled: 672703 tps, 97.13% utilization 50/90/99% latency:244.31 484.205 1028.41 No-cache copy enabled: 702113 tps, 96.16% utilization, 50/90/99% latency 238.56 467.56 956.955 Using 14000 byte request and response sizes demonstrate the effects more dramatically: No-cache copy disabled: 79571 tps, 34.34 %utlization 50/90/95% latency 1584.46 2319.59 5001.76 No-cache copy enabled: 83856 tps, 34.81% utilization 50/90/95% latency 2508.42 2622.62 2735.88 Note especially the effect on latency tail (95th percentile). This seems to provide a nice performance improvement and is consistent in the tests I ran. Presumably, this would provide the greatest benfits in the presence of an application workload stressing the cache and a lot of transmit data happening. Signed-off-by: NTom Herbert <therbert@google.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 stephen hemminger 提交于
Apply restrictions on STP parameters based 802.1D 1998 standard. * Fixes missing locking in set path cost ioctl * Uses common code for both ioctl and sysfs This is based on an earlier patch Sasikanth V but with overhaul. Note: 1. It does NOT enforce the restriction on the relationship max_age and forward delay or hello time because in existing implementation these are set as independant operations. 2. If STP is disabled, there is no restriction on forward delay 3. No restriction on holding time because users use Linux code to act as hub or be sticky. 4. Although standard allow 0-255, Linux only allows 0-63 for port priority because more bits are reserved for port number. Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 stephen hemminger 提交于
Add netlink device ops to allow creating bridge device via netlink. This works in a manner similar to vlan, macvlan and bonding. Example: # ip link add link dev br0 type bridge # ip link del dev br0 The change required rearranging initializtion code to deal with being called by create link. Most of the initialization happens in br_dev_setup, but allocation of stats is done in ndo_init callback to deal with allocation failure. Sysfs setup has to wait until after the network device kobject is registered. Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 stephen hemminger 提交于
Use RTM_NEWNEIGH and RTM_DELNEIGH to allow updating of entries in bridge forwarding table. This allows manipulating static entries which is not possible with existing tools. Example (using bridge extensions to iproute2) # br fdb add 00:02:03:04:05:06 dev eth0 Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 stephen hemminger 提交于
This allows applications to query and monitor bridge forwarding table in the same method used for neighbor table. The forward table entries are returned in same structure format as used by the ioctl. If more information is desired in future, the netlink method is extensible. Example (using bridge extensions to iproute2) # br monitor Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 stephen hemminger 提交于
In some cases, look up of forward database entry is done with RCU; and for others no RCU is needed because of locking. Split the two cases into two differnt loops (and take off inline). Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 stephen hemminger 提交于
Adds tracking the last used time in forwarding table. Rename ageing_timer to updated to better describe it. Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 stephen hemminger 提交于
Later patch provides ability to create non-local static entry. To make this easier move the updating of the flag values to after the code that creates entry. Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Boris Ostrovsky 提交于
Make sure dst_release() is not called with error pointer. This is similar to commit 4910ac6c ("ipv4: Don't ip_rt_put() an error pointer in RAW sockets."). Signed-off-by: NBoris Ostrovsky <boris.ostrovsky@amd.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Helmut Schaa 提交于
Cleaning the ieee80211_rx_data.flags field here is wrong, instead the flags should be valid accross processing the frame on different interfaces. Fix this by removing the incorrect flags=0 assignment. Introduced in commit 554891e6 (mac80211: move packet flags into packet). Signed-off-by: NHelmut Schaa <helmut.schaa@googlemail.com> Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
-
由 stephen hemminger 提交于
This is an implementation of the Quick Fair Queue scheduler developed by Fabio Checconi. The same algorithm is already implemented in ipfw in FreeBSD. Fabio had an earlier version developed on Linux, I just cleaned it up. Thanks to Eric Dumazet for testing this under load. Signed-off-by: NStephen Hemminger <shemminger@vyatta.com> Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 04 4月, 2011 8 次提交
-
-
由 Florian Westphal 提交于
--ctdir ORIGINAL matches REPLY packets, and vv: userspace sets "invert_flags &= ~XT_CONNTRACK_DIRECTION" in ORIGINAL case. Thus: (CTINFO2DIR(ctinfo) == IP_CT_DIR_ORIGINAL) ^ !!(info->invert_flags & XT_CONNTRACK_DIRECTION)) yields "1 ^ 0", which is true -> returns false. Reproducer: iptables -I OUTPUT 1 -p tcp --syn -m conntrack --ctdir ORIGINAL Signed-off-by: NFlorian Westphal <fwestphal@astaro.com> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
由 Florian Westphal 提交于
This avoids pulling in the ipv6 module when using (ipv4-only) iptables -m addrtype. Signed-off-by: NFlorian Westphal <fw@strlen.de> Acked-by: NDavid S. Miller <davem@davemloft.net> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
由 Florian Westphal 提交于
ipv6 fib lookup can set RT6_LOOKUP_F_IFACE flag to restrict search to an interface, but this flag cannot be set via struct flowi. Also, it cannot be set via ip6_route_output: this function uses the passed sock struct to determine if this flag is required (by testing for nonzero sk_bound_dev_if). Work around this by passing in an artificial struct sk in case 'strict' argument is true. This is required to replace the rt6_lookup call in xt_addrtype.c with nf_afinfo->route(). Signed-off-by: NFlorian Westphal <fw@strlen.de> Acked-by: NDavid S. Miller <davem@davemloft.net> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
由 Florian Westphal 提交于
This is required to eventually replace the rt6_lookup call in xt_addrtype.c with nf_afinfo->route(). Signed-off-by: NFlorian Westphal <fw@strlen.de> Acked-by: NDavid S. Miller <davem@davemloft.net> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
由 Hans Schillstrom 提交于
ipvsadm -ln --daemon will trigger a Null pointer exception because ip_vs_genl_dump_daemons() uses skb_net() instead of skb_sknet(). To prevent others from NULL ptr a check is made in ip_vs.h skb_net(). Signed-off-by: NHans Schillstrom <hans.schillstrom@ericsson.com> Signed-off-by: NSimon Horman <horms@verge.net.au> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
由 David Sterba 提交于
Static analyzer of clang found a dead store which appears to be a bug in reading count of items in SEQOF field, only the lower byte of word is stored. This may lead to corrupted read and communication shutdown. The bug has been in the module since it's first inclusion into linux kernel. [Patrick: the bug is real, but without practical consequence since the largest amount of sequence-of members we parse is 30.] Signed-off-by: NDavid Sterba <dsterba@suse.cz> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
由 Jozsef Kadlecsik 提交于
The timeout variant of the list:set type must reference the member sets. However, its garbage collector runs at timer interrupt so the mutex protection of the references is a no go. Therefore the reference protection is converted to rwlock. Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
由 Jozsef Kadlecsik 提交于
- the timeout value was actually not set - the garbage collector was broken The variant is fixed, the tests to the ipset testsuite are added. Signed-off-by: NJozsef Kadlecsik <kadlec@blackhole.kfki.hu> Signed-off-by: NPatrick McHardy <kaber@trash.net>
-
- 03 4月, 2011 2 次提交
-
-
由 Michał Mirosław 提交于
Note: get_flags was actually broken, because it should return the flags capped with vlan_features. This is now done implicitly by limiting netdev->hw_features. RX checksumming offload control is (and was) broken, as there was no way before to say whether it's done for tagged packets. Signed-off-by: NMichał Mirosław <mirq-linux@rere.qmqm.pl> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Michał Mirosław 提交于
Issue FEAT_CHANGE notification when features are changed by netdev_update_features(). This will allow changes made by extra constraints on e.g. MTU change to be properly propagated like changes via ethtool. Signed-off-by: NMichał Mirosław <mirq-linux@rere.qmqm.pl> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 02 4月, 2011 2 次提交
-
-
由 Ilpo Järvinen 提交于
All callers are prepared for alloc failures anyway, so this error can safely be boomeranged to the callers domain without super bad consequences. ...At worst the connection might go into a state where each RTO tries to (unsuccessfully) re-fragment with such a mis-sized value and eventually dies. Signed-off-by: NIlpo Järvinen <ilpo.jarvinen@helsinki.fi> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Wei Yongjun 提交于
Sometime the ASCONF_ACK parameters can equal to the fourfold of ASCONF parameters, this only happend in some special case: ASCONF parameter is : Unrecognized Parameter (4 bytes) ASCONF_ACK parameter should be: Error Cause Indication parameter (8 bytes header) + Error Cause (4 bytes header) + Unrecognized Parameter (4bytes) Four 4bytes Unrecognized Parameters in ASCONF chunk will cause panic. Pid: 0, comm: swapper Not tainted 2.6.38-next+ #22 Bochs Bochs EIP: 0060:[<c0717eae>] EFLAGS: 00010246 CPU: 0 EIP is at skb_put+0x60/0x70 EAX: 00000077 EBX: c09060e2 ECX: dec1dc30 EDX: c09469c0 ESI: 00000000 EDI: de3c8d40 EBP: dec1dc58 ESP: dec1dc2c DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 Process swapper (pid: 0, ti=dec1c000 task=c09aef20 task.ti=c0980000) Stack: c09469c0 e1894fa4 00000044 00000004 de3c8d00 de3c8d00 de3c8d44 de3c8d40 c09060e2 de25dd80 de3c8d40 dec1dc7c e1894fa4 dec1dcb0 00000040 00000004 00000000 00000800 00000004 00000004 dec1dce0 e1895a2b dec1dcb4 de25d960 Call Trace: [<e1894fa4>] ? sctp_addto_chunk+0x4e/0x89 [sctp] [<e1894fa4>] sctp_addto_chunk+0x4e/0x89 [sctp] [<e1895a2b>] sctp_process_asconf+0x32f/0x3d1 [sctp] [<e188d554>] sctp_sf_do_asconf+0xf8/0x173 [sctp] [<e1890b02>] sctp_do_sm+0xb8/0x159 [sctp] [<e18a2248>] ? sctp_cname+0x0/0x52 [sctp] [<e189392d>] sctp_assoc_bh_rcv+0xac/0xe3 [sctp] [<e1897d76>] sctp_inq_push+0x2d/0x30 [sctp] [<e18a21b2>] sctp_rcv+0x7a7/0x83d [sctp] [<c077a95c>] ? ipv4_confirm+0x118/0x125 [<c073a970>] ? nf_iterate+0x34/0x62 [<c074789d>] ? ip_local_deliver_finish+0x0/0x194 [<c074789d>] ? ip_local_deliver_finish+0x0/0x194 [<c0747992>] ip_local_deliver_finish+0xf5/0x194 [<c074789d>] ? ip_local_deliver_finish+0x0/0x194 [<c0747a6e>] NF_HOOK.clone.1+0x3d/0x44 [<c0747ab3>] ip_local_deliver+0x3e/0x44 [<c074789d>] ? ip_local_deliver_finish+0x0/0x194 [<c074775c>] ip_rcv_finish+0x29f/0x2c7 [<c07474bd>] ? ip_rcv_finish+0x0/0x2c7 [<c0747a6e>] NF_HOOK.clone.1+0x3d/0x44 [<c0747cae>] ip_rcv+0x1f5/0x233 [<c07474bd>] ? ip_rcv_finish+0x0/0x2c7 [<c071dce3>] __netif_receive_skb+0x310/0x336 [<c07221f3>] netif_receive_skb+0x4b/0x51 [<e0a4ed3d>] cp_rx_poll+0x1e7/0x29c [8139cp] [<c072275e>] net_rx_action+0x65/0x13a [<c0445a54>] __do_softirq+0xa1/0x149 [<c04459b3>] ? __do_softirq+0x0/0x149 <IRQ> [<c0445891>] ? irq_exit+0x37/0x72 [<c040a7e9>] ? do_IRQ+0x81/0x95 [<c07b3670>] ? common_interrupt+0x30/0x38 [<c0428058>] ? native_safe_halt+0xa/0xc [<c040f5d7>] ? default_idle+0x58/0x92 [<c0408fb0>] ? cpu_idle+0x96/0xb2 [<c0797989>] ? rest_init+0x5d/0x5f [<c09fd90c>] ? start_kernel+0x34b/0x350 [<c09fd0cb>] ? i386_start_kernel+0xba/0xc1 Signed-off-by: NWei Yongjun <yjwei@cn.fujitsu.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 01 4月, 2011 1 次提交
-
-
由 David S. Miller 提交于
Commit 60d9f461 ("appletalk: remove the BKL") added a dereference of "sk" before checking for NULL in atalk_release(). Guard the code block completely, rather than partially, with the NULL check. Reported-by: NDave Jones <davej@redhat.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 31 3月, 2011 9 次提交
-
-
由 David S. Miller 提交于
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 David S. Miller 提交于
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 David S. Miller 提交于
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 David S. Miller 提交于
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 David S. Miller 提交于
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Eric Dumazet 提交于
Add __rcu annotations and lockdep checks. Add const qualifiers node_parent() and node_parent_rcu() can use rcu_dereference_index_check() Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 David S. Miller 提交于
Like DCCP and other similar pieces of code, there are mechanisms here to try allocating smaller hash tables if the allocation fails. So pass in __GFP_NOWARN like the others do instead of emitting a scary message. Reported-by: NDave Jones <davej@redhat.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Eric Dumazet 提交于
Daniel J Blueman reported a lockdep splat in trie_firstleaf(), caused by RTNL being not locked before a call to fib_table_flush() Reported-by: NDaniel J Blueman <daniel.blueman@gmail.com> Signed-off-by: NEric Dumazet <eric.dumazet@gmail.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Philip A. Prindeville 提交于
Don't flap VCs when carrier state changes; higher-level protocols can detect loss of connectivity and act accordingly. This is more consistent with how other network interfaces work. We no longer use release_vccs() so we can delete it. release_vccs() was duplicated from net/atm/common.c; make the corresponding function exported, since other code duplicates it and could leverage it if it were public. Signed-off-by: NPhilip A. Prindeville <philipp@redfish-solutions.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 30 3月, 2011 1 次提交
-
-
由 Daniel Lezcano 提交于
In case the device where is coming from the packet has TSO enabled, we should not check the mtu size value as this one could be bigger than the expected value. This is the case for the macvlan driver when the lower device has TSO enabled. The macvlan inherit this feature and forward the packets without fragmenting them. Then the packets go through dev_forward_skb and are dropped. This patch fix this by checking TSO is not enabled when we want to check the mtu size. Signed-off-by: NDaniel Lezcano <daniel.lezcano@free.fr> Acked-by: N"Eric W. Biederman" <ebiederm@xmission.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-