aboutsummaryrefslogtreecommitdiff
path: root/net
AgeCommit message (Collapse)AuthorFilesLines
2013-10-29net, mc: fix the incorrect comments in two mc-related functionsZhi Yong Wu1-2/+2
Signed-off-by: Zhi Yong Wu <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-29net, iovec: fix the incorrect comment in memcpy_fromiovecend()Zhi Yong Wu1-1/+1
Signed-off-by: Zhi Yong Wu <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-29net, datagram: fix the incorrect comment in zerocopy_sg_from_iovec()Zhi Yong Wu1-1/+1
Signed-off-by: Zhi Yong Wu <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-29ipv4: fix DO and PROBE pmtu mode regarding local fragmentation with UFO/CORKHannes Frederic Sowa1-4/+8
UFO as well as UDP_CORK do not respect IP_PMTUDISC_DO and IP_PMTUDISC_PROBE well enough. UFO enabled packet delivery just appends all frags to the cork and hands it over to the network card. So we just deliver non-DF udp fragments (DF-flag may get overwritten by hardware or virtual UFO enabled interface). UDP_CORK does enqueue the data until the cork is disengaged. At this point it sets the correct IP_DF and local_df flags and hands it over to ip_fragment which in this case will generate an icmp error which gets appended to the error socket queue. This is not reflected in the syscall error (of course, if UFO is enabled this also won't happen). Improve this by checking the pmtudisc flags before appending data to the socket and if we still can fit all data in one packet when IP_PMTUDISC_DO or IP_PMTUDISC_PROBE is set, only then proceed. We use (mtu-fragheaderlen) to check for the maximum length because we ensure not to generate a fragment and non-fragmented data does not need to have its length aligned on 64 bit boundaries. Also the passed in ip_options are already aligned correctly. Maybe, we can relax some other checks around ip_fragment. This needs more research. Signed-off-by: Hannes Frederic Sowa <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-29tcp: gso: fix truesize trackingEric Dumazet1-8/+5
commit 6ff50cd55545 ("tcp: gso: do not generate out of order packets") had an heuristic that can trigger a warning in skb_try_coalesce(), because skb->truesize of the gso segments were exactly set to mss. This breaks the requirement that skb->truesize >= skb->len + truesizeof(struct sk_buff); It can trivially be reproduced by : ifconfig lo mtu 1500 ethtool -K lo tso off netperf As the skbs are looped into the TCP networking stack, skb_try_coalesce() warns us of these skb under-estimating their truesize. Signed-off-by: Eric Dumazet <[email protected]> Reported-by: Alexei Starovoitov <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-28ipv6: Remove privacy config option.David S. Miller2-56/+3
The code for privacy extentions is very mature, and making it configurable only gives marginal memory/code savings in exchange for obfuscation and hard to read code via CPP ifdef'ery. Signed-off-by: David S. Miller <[email protected]>
2013-10-286lowpan: remove unnecessary breakAlexander Aring1-1/+0
Signed-off-by: Alexander Aring <[email protected]> Reviewed-by: Werner Almesberger <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-286lowpan: remove skb->dev assignmentAlexander Aring1-1/+0
This patch removes the assignment of skb->dev. We don't need it here because we use the netdev_alloc_skb_ip_align function which already sets the skb->dev. Signed-off-by: Alexander Aring <[email protected]> Reviewed-by: Werner Almesberger <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-286lowpan: use netdev_alloc_skb instead dev_alloc_skbAlexander Aring1-2/+2
This patch uses the netdev_alloc_skb instead dev_alloc_skb function and drops the seperate assignment to skb->dev. Signed-off-by: Alexander Aring <[email protected]> Reviewed-by: Werner Almesberger <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-286lowpan: remove unnecessary check on err >= 0Alexander Aring1-1/+1
The err variable can only be zero in this case. Signed-off-by: Alexander Aring <[email protected]> Reviewed-by: Werner Almesberger <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-286lowpan: remove unnecessary ret variableAlexander Aring1-4/+2
Signed-off-by: Alexander Aring <[email protected]> Reviewed-by: Werner Almesberger <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-28SUNRPC: Fix buffer overflow checking in gss_encode_v0_msg/gss_encode_v1_msgTrond Myklebust1-19/+37
In gss_encode_v1_msg, it is pointless to BUG() after the overflow has happened. Replace the existing sprintf()-based code with scnprintf(), and warn if an overflow is ever triggered. In gss_encode_v0_msg, replace the runtime BUG_ON() with an appropriate compile-time BUILD_BUG_ON. Reported-by: Bruce Fields <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2013-10-28SUNRPC: gss_alloc_msg - choose _either_ a v0 message or a v1 messageTrond Myklebust1-0/+1
Add the missing 'break' to ensure that we don't corrupt a legacy 'v0' type message by appending the 'v1'. Cc: Bruce Fields <[email protected]> Cc: [email protected] Signed-off-by: Trond Myklebust <[email protected]>
2013-10-28SUNRPC: remove an unnecessary if statementwangweidong1-3/+1
If req allocated failed just goto out_free, no need to check the 'i < num_prealloc'. There is just code simplification, no functional changes. Signed-off-by: Wang Weidong <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2013-10-28sunrpc: comment typo fixJ. Bruce Fields1-2/+2
Signed-off-by: J. Bruce Fields <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2013-10-28SUNRPC: Add correct rcu_dereference annotation in rpc_clnt_set_transportTrond Myklebust1-1/+2
rpc_clnt_set_transport should use rcu_derefence_protected(), as it is only safe to be called with the rpc_clnt::cl_lock held. Cc: Chuck Lever <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2013-10-28SUNRPC: Add a helper to switch the transport of an rpc_clntTrond Myklebust1-7/+100
Add an RPC client API to redirect an rpc_clnt's transport from a source server to a destination server during a migration event. Signed-off-by: Trond Myklebust <[email protected]> [ cel: forward ported to 3.12 ] Signed-off-by: Chuck Lever <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2013-10-28SUNRPC: Modify synopsis of rpc_client_register()Chuck Lever1-6/+7
The rpc_client_register() helper was added in commit e73f4cc0, "SUNRPC: split client creation routine into setup and registration," Mon Jun 24 11:52:52 2013. In a subsequent patch, I'd like to invoke rpc_client_register() from a context where a struct rpc_create_args is not available. Signed-off-by: Chuck Lever <[email protected]> Signed-off-by: Trond Myklebust <[email protected]>
2013-10-28netfilter: bridge: nf_tables: add filter chain typePablo Neira Ayuso1-2/+39
This patch adds the filter chain type which is required to create filter chains in the bridge family from userspace. Signed-off-by: Pablo Neira Ayuso <[email protected]>
2013-10-28netfilter: nft_nat: Fix endianness issue reported by sparseTomasz Bursztyka1-4/+8
This patch fixes this: CHECK net/netfilter/nft_nat.c net/netfilter/nft_nat.c:50:43: warning: incorrect type in assignment (different base types) net/netfilter/nft_nat.c:50:43: expected restricted __be32 [addressable] [usertype] ip net/netfilter/nft_nat.c:50:43: got unsigned int [unsigned] [usertype] <noident> net/netfilter/nft_nat.c:51:43: warning: incorrect type in assignment (different base types) net/netfilter/nft_nat.c:51:43: expected restricted __be32 [addressable] [usertype] ip net/netfilter/nft_nat.c:51:43: got unsigned int [unsigned] [usertype] <noident> net/netfilter/nft_nat.c:65:37: warning: incorrect type in assignment (different base types) net/netfilter/nft_nat.c:65:37: expected restricted __be16 [addressable] [assigned] [usertype] all net/netfilter/nft_nat.c:65:37: got unsigned int [unsigned] <noident> net/netfilter/nft_nat.c:66:37: warning: incorrect type in assignment (different base types) net/netfilter/nft_nat.c:66:37: expected restricted __be16 [addressable] [assigned] [usertype] all net/netfilter/nft_nat.c:66:37: got unsigned int [unsigned] <noident> Signed-off-by: Tomasz Bursztyka <[email protected]> Signed-off-by: Pablo Neira Ayuso <[email protected]>
2013-10-28netfilter: bridge: fix nf_tables bridge dependencies with main corePablo Neira Ayuso1-0/+1
when CONFIG_NF_TABLES[_MODULE] is not enabled, but CONFIG_NF_TABLES_BRIDGE is enabled: net/bridge/netfilter/nf_tables_bridge.c: In function 'nf_tables_bridge_init_net': net/bridge/netfilter/nf_tables_bridge.c:24:5: error: 'struct net' has no member named 'nft' net/bridge/netfilter/nf_tables_bridge.c:25:9: error: 'struct net' has no member named 'nft' net/bridge/netfilter/nf_tables_bridge.c:28:2: error: 'struct net' has no member named 'nft' net/bridge/netfilter/nf_tables_bridge.c:30:34: error: 'struct net' has no member named 'nft' net/bridge/netfilter/nf_tables_bridge.c:35:11: error: 'struct net' has no member named 'nft' net/bridge/netfilter/nf_tables_bridge.c: In function 'nf_tables_bridge_exit_net': net/bridge/netfilter/nf_tables_bridge.c:41:27: error: 'struct net' has no member named 'nft' net/bridge/netfilter/nf_tables_bridge.c:42:11: error: 'struct net' has no member named 'nft' Reported-by: Randy Dunlap <[email protected]> Reported-by: kbuild test robot <[email protected]> Signed-off-by: Pablo Neira Ayuso <[email protected]>
2013-10-28nl80211: fix channel switch parsingAndrei Otcheretianski1-3/+1
The nl80211 attribute NL80211_ATTR_CSA_C_OFF_BEACON should be nested inside NL80211_ATTR_CSA_IES, but commit ee4bc9e75811d2c0cb5f2a2fc5b5 ("nl80211: enable IBSS support for channel switch announcements") added a check in the outer message attributes. Fix channel switch calls by removing the erroneus condition. Signed-off-by: Andrei Otcheretianski <[email protected]> [reword commit message] Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: fix uninitialized variableMichal Kazior1-1/+1
CSA completion could call in a driver bss_info_changed() with a garbled `changed` flag leading to all sorts of problems. Signed-off-by: Michal Kazior <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: process mesh channel switching using beaconChun-Yeow Yeoh2-9/+136
Trigger the mesh channel switching procedure if the mesh STA happens to miss the CSA action frame but able to receive the beacon containing the CSA and MCSP elements from its peer mesh STAs. Signed-off-by: Chun-Yeow Yeoh <[email protected]> [fix locking in ieee80211_mesh_process_chnswitch()] Signed-off-by: Johannes Berg <[email protected]>
2013-10-28{nl,cfg,mac}80211: implement mesh channel switch userspace APIChun-Yeow Yeoh6-3/+139
Implement the required procedures for mesh channel switching as defined in the IEEE Std 802.11-2012 section 10.9.8.4.3 and also handle the CSA and MCSP elements as followed: * Add the function for updating the beacon and probe response frames with CSA and MCSP elements during the period of switching to the new channel. Both CSA and MCSP elements must be included in beacon and probe response frames until the intended channel switch time. * The ifmsh->csa_settings is set to NULL and the CSA and MCSP elements will then be removed from the beacon or probe response frames once the new channel is switched to. Signed-off-by: Chun-Yeow Yeoh <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28{nl,cfg,mac}80211: enable the triggering of CSA frame in meshChun-Yeow Yeoh5-54/+111
Allow the triggering of CSA frame using mesh interface. The rules are more or less same with IBSS, such as not allowed to change between the band and channel width has to be same from the previous mode. Also, move the ieee80211_send_action_csa to a common space so that it can be re-used by mesh interface. Signed-off-by: Chun-Yeow Yeoh <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: process the CSA frame for mesh accordinglyChun-Yeow Yeoh5-3/+114
Process the CSA frame according to the procedures define in IEEE Std 802.11-2012 section 10.9.8.4.3 as follow: * The mesh channel switch parameters element (MCSP) must be availabe. * If the MCSP's TTL is 1, drop the frame but still process the CSA. * If the MCSP's precedence value is less than or equal to the current precedence value, drop the frame and do not process the CSA. * The CSA frame is forwarded after TTL is decremented by 1 and the initiator field is set to 0. Transmit restrict field and others are maintained as is. * No beacon or probe response frame are handled here. Also, introduce the debug message used for mesh CSA purpose. Signed-off-by: Chun-Yeow Yeoh <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: refactor the parsing of chan switch ieChun-Yeow Yeoh4-41/+48
Refactor the channel switch IE parsing to reduce the number of function parameters. Signed-off-by: Chun-Yeow Yeoh <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: add ieee80211_tx_prepare_skb() helper functionFelix Fietkau1-0/+29
This can be used by a driver to prepare skbs for transmission, which were obtained via functions such as ieee80211_probereq_get or ieee80211_nullfunc_get. This is useful for drivers that want to send those frames directly, but need rate control information to be prepared first. Signed-off-by: Felix Fietkau <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28cfg80211: export reg_initiator_name()Luis R. Rodriguez1-10/+16
Drivers can now use this to parse the regulatory request and be more verbose when needed. Signed-off-by: Luis R. Rodriguez <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: fixes for mesh powersave logicMarco Porsch3-18/+22
This patch fixes errors in the mesh powersave logic which cause that remote peers do not get peer power mode change notifications and mesh peer service periods (MPSPs) got stuck. When closing a peer link, set the (now invalid) peer-specific power mode to 'unknown'. Avoid overhead when local power mode is unchanged. Reliably clear MPSP flags on peering status update. Avoid MPSP flags getting stuck by not requesting a further MPSP ownership if we already are an MPSP owner. Signed-off-by: Marco Porsch <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: Remove check for offchannel state when waking netdev queuesSeth Forshee1-3/+0
6c17b77b67587b9f9e3070fb89fe98cef3187131 ensures that a device's mac80211 queues will remain stopped while offchannel. Since the vif can no longer be offchannel when the queues wake it's not necessary to check for this before waking its netdev queues. Signed-off-by: Seth Forshee <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28net: rfkill: gpio: add ACPI supportHeikki Krogerus1-1/+30
Including ACPI ID for Broadcom GPS receiver BCM4752. Signed-off-by: Heikki Krogerus <[email protected]> Acked-by: Rhyland Klein <[email protected]> Reviewed-by: Mika Westerberg <[email protected]> Acked-by: Rafael J. Wysocki <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28net: rfkill: gpio: prepare for DT and ACPI supportHeikki Krogerus1-41/+51
This will add the relevant values like the gpios and the type in rfkill_gpio_platform_data to the rfkill_gpio_data structure. It will allow those values to be easily picked from DT and ACPI tables later. Signed-off-by: Heikki Krogerus <[email protected]> Acked-by: Rhyland Klein <[email protected]> Reviewed-by: Mika Westerberg <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28net: rfkill: gpio: spinlock-safe GPIO accessHeikki Krogerus1-8/+8
This sets the direction of the gpio once when it's requested, and uses the spinlock-safe gpio_set_state() to change the state. Signed-off-by: Heikki Krogerus <[email protected]> Acked-by: Rhyland Klein <[email protected]> Reviewed-by: Mika Westerberg <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28net: rfkill: gpio: clean up clock handlingHeikki Krogerus2-28/+9
Use a simple flag to see the state of the clock, and make the clock available even without a name. Also, get rid of HAVE_CLK dependency. Signed-off-by: Heikki Krogerus <[email protected]> Acked-by: Rhyland Klein <[email protected]> Reviewed-by: Mika Westerberg <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28net: rfkill: gpio: convert to resource managed allocationHeikki Krogerus1-55/+20
And remove now unneeded resource freeing. Signed-off-by: Heikki Krogerus <[email protected]> Acked-by: Rhyland Klein <[email protected]> Reviewed-by: Mika Westerberg <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: enable DFS for IBSS modeSimon Wunderlich2-5/+45
Allow changing to DFS channels if the channel is available for beaconing and userspace controls DFS operation. Channel switch announcement from other stations on DFS channels will be interpreted as radar event. These channels will then be marked as unvailable. Signed-off-by: Simon Wunderlich <[email protected]> Signed-off-by: Mathias Kretschmer <[email protected]> Signed-off-by: Johannes Berg <[email protected]>
2013-10-28nl80211/cfg80211: enable DFS for IBSS modeSimon Wunderlich4-11/+38
To use DFS in IBSS mode, userspace is required to react to radar events. It can inform nl80211 that it is capable of doing so by adding a NL80211_ATTR_HANDLE_DFS attribute when joining the IBSS. This attribute is supplied to let the kernelspace know that the userspace application can and will handle radar events, e.g. by intiating channel switches to a valid channel. DFS channels may only be used if this attribute is supplied and the driver supports it. Driver support will be checked even if a channel without DFS will be initially joined, as a DFS channel may be chosen later. Signed-off-by: Simon Wunderlich <[email protected]> Signed-off-by: Mathias Kretschmer <[email protected]> [fix attribute name in commit message] Signed-off-by: Johannes Berg <[email protected]>
2013-10-28mac80211: implement SMPS for APEmmanuel Grumbach10-38/+289
When the driver requests to move to STATIC or DYNAMIC SMPS, we send an action frame to each associated station and reconfigure the channel context / driver. Of course, non-MIMO stations are ignored. The beacon isn't updated. The association response will include the original capabilities. Stations that associate while in non-OFF SMPS mode will get an action frame right after association to inform them about our current state. Note that we wait until the end of the EAPOL. Sending an action frame before the EAPOL is finished can be an issue for a few clients. Clients aren't likely to send EAPOL frames in MIMO anyway. When the SMPS configuration gets more permissive (e.g. STATIC -> OFF), we don't wake up stations that are asleep We remember that they don't know about the change and send the action frame when they wake up. When the SMPS configuration gets more restrictive (e.g. OFF -> STATIC), we set the TIM bit for every sleeping STA. uAPSD stations might send MIMO until they poll the action frame, but this is for a short period of time. Signed-off-by: Emmanuel Grumbach <[email protected]> [fix vht streams loop, initialisation] Signed-off-by: Johannes Berg <[email protected]>
2013-10-28net: ipvs: sctp: add missing verdict assignments in sctp_conn_scheduleDaniel Borkmann1-2/+7
If skb_header_pointer() fails, we need to assign a verdict, that is NF_DROP in this case, otherwise, we would leave the verdict from conn_schedule() uninitialized when returning. Signed-off-by: Daniel Borkmann <[email protected]> Acked-by: Jesper Dangaard Brouer <[email protected]> Acked-by: Neil Horman <[email protected]> Acked-by: Julian Anastasov <[email protected]> Signed-off-by: Simon Horman <[email protected]>
2013-10-28xfrm: Increase the garbage collector thresholdSteffen Klassert2-2/+2
With the removal of the routing cache, we lost the option to tweak the garbage collector threshold along with the maximum routing cache size. So git commit 703fb94ec ("xfrm: Fix the gc threshold value for ipv4") moved back to a static threshold. It turned out that the current threshold before we start garbage collecting is much to small for some workloads, so increase it from 1024 to 32768. This means that we start the garbage collector if we have more than 32768 dst entries in the system and refuse new allocations if we are above 65536. Reported-by: Wolfgang Walter <[email protected]> Signed-off-by: Steffen Klassert <[email protected]>
2013-10-28sctp: merge two if statements to onewangweidong1-9/+5
Two if statements do the same work, we can merge them to one. And fix some typos. There is just code simplification, no functional changes. Signed-off-by: Wang Weidong <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-28sctp: remove the repeat initialize with 0wangweidong1-21/+8
kmem_cache_zalloc had set the allocated memory to zero. I think no need to initialize with 0. And move the comments to the function begin. Acked-by: Vlad Yasevich <[email protected]> Signed-off-by: Wang Weidong <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-28sctp: fix some comments in chunk.c and associola.cwangweidong2-3/+3
fix some typos Acked-by: Vlad Yasevich <[email protected]> Signed-off-by: Wang Weidong <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-28inet: restore gso for vxlanEric Dumazet1-8/+7
Alexei reported a performance regression on vxlan, caused by commit 3347c9602955 "ipv4: gso: make inet_gso_segment() stackable" GSO vxlan packets were not properly segmented, adding IP fragments while they were not expected. Rename 'bool tunnel' to 'bool encap', and add a new boolean to express the fact that UDP should be fragmented. This fragmentation is triggered by skb->encapsulation being set. Remove a "skb->encapsulation = 1" added in above commit, as its not needed, as frags inherit skb->frag from original GSO skb. Reported-by: Alexei Starovoitov <[email protected]> Signed-off-by: Eric Dumazet <[email protected]> Tested-by: Alexei Starovoitov <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-28pkt_sched: fq: clear time_next_packet for reused flowsEric Dumazet1-0/+1
When a socket is freed/reallocated, we need to clear time_next_packet or else we can inherit a prior value and delay first packets of the new flow. Signed-off-by: Eric Dumazet <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-27tcp: do not rearm RTO when future data are sackedYuchung Cheng1-3/+10
Patch ed08495c3 "tcp: use RTT from SACK for RTO" always re-arms RTO upon obtaining a RTT sample from newly sacked data. But technically RTO should only be re-armed when the data sent before the last (re)transmission of write queue head are (s)acked. Otherwise the RTO may continue to extend during loss recovery on data sent in the future. Note that RTTs from ACK or timestamps do not have this problem, as the RTT source must be from data sent before. The new RTO re-arm policy is 1) Always re-arm RTO if SND.UNA is advanced 2) Re-arm RTO if sack RTT is available, provided the sacked data was sent before the last time write_queue_head was sent. Signed-off-by: Larry Brakmo <[email protected]> Signed-off-by: Yuchung Cheng <[email protected]> Acked-by: Neal Cardwell <[email protected]> Acked-by: Eric Dumazet <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-27tcp: only take RTT from timestamps if new data is ackedYuchung Cheng1-1/+2
Patch ed08495c3 "tcp: use RTT from SACK for RTO" has a bug that it does not check if the ACK acknowledge new data before taking the RTT sample from TCP timestamps. This patch adds the check back as required by the RFC. Signed-off-by: Yuchung Cheng <[email protected]> Acked-by: Neal Cardwell <[email protected]> Acked-by: Eric Dumazet <[email protected]> Signed-off-by: David S. Miller <[email protected]>
2013-10-27tcp: fix SYNACK RTT estimation in Fast OpenYuchung Cheng1-5/+13
tp->lsndtime may not always be the SYNACK timestamp if a passive Fast Open socket sends data before handshake completes. And if the remote acknowledges both the data and the SYNACK, the RTT sample is already taken in tcp_ack(), so no need to call tcp_update_ack_rtt() in tcp_synack_rtt_meas() aagain. Signed-off-by: Yuchung Cheng <[email protected]> Acked-by: Neal Cardwell <[email protected]> Acked-by: Eric Dumazet <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.