summaryrefslogtreecommitdiff
path: root/net
AgeCommit message (Collapse)Author
2010-08-28econet: kill unnecessary spin_lock_init()Akinobu Mita
The spinlock aun_queue_lock is initialized statically. It is unnecessary to initialize by spin_lock_init() at module load time. This is detected by the semantic patch. // <smpl> @def@ declarer name DEFINE_SPINLOCK; identifier spinlock; @@ DEFINE_SPINLOCK(spinlock); @@ identifier def.spinlock; @@ - spin_lock_init(&spinlock); // </smpl> Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com> Cc: Julia Lawall <julia@diku.dk> Cc: "David S. Miller" <davem@davemloft.net> Cc: netdev@vger.kernel.org Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-27net/ipv4: Eliminate kstrdup memory leakJulia Lawall
The string clone is only used as a temporary copy of the argument val within the while loop, and so it should be freed before leaving the function. The call to strsep, however, modifies clone, so a pointer to the front of the string is kept in saved_clone, to make it possible to free it. The sematic match that finds this problem is as follows: (http://coccinelle.lip6.fr/) // <smpl> @r exists@ local idexpression x; expression E; identifier l; statement S; @@ *x= \(kasprintf\|kstrdup\)(...); ... if (x == NULL) S ... when != kfree(x) when != E = x if (...) { <... when != kfree(x) * goto l; ...> * return ...; } // </smpl> Signed-off-by: Julia Lawall <julia@diku.dk> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-27mac80211: fix offchannel queue stopJohannes Berg
Somebody noticed this problem, and I outlined to them how to fix it, but haven't heard back from them. So while I was adding the state field I figured I could use it to fix it. The problem, as I understand it, is that when we go offchannel while the driver has a queue stopped, the driver will likely start draining the queue and then enable it while offchannel. This in turn will enable the interface queue, and that leads to transmitting data frames on the wrong channel. Fix this by keeping track of offchannel status per interface, and not enabling the interface queues on interfaces that are offchannel when the driver enables a queue. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: support runtime interface type changesJohannes Berg
Add support to mac80211 for changing the interface type even when the interface is UP, if the driver supports it. To achieve this * add a new driver callback for switching, * split some of the interface up/down code out into new functions (do_open/do_stop), and * maintain an own __SDATA_RUNNING bit that will not be set during interface type, so that any other code doesn't use the interface. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: split out concurrent vif checksJohannes Berg
Split the concurrent virtual interface checks into a new function that can be used to check for any given new interface type. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: simplify zero address checksJohannes Berg
The libertas_tf special code for zero addresses is a bit too complex, it compares against a stack value instead of using is_zero_ether_addr() and tries to update all interfaces even if just the one that's being brought up needs to be changed. Additionally, the repeated check for a valid MAC address need only be done if we actually changed it on the fly. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: switch to ieee80211_sdata_runningJohannes Berg
Since the introduction of ieee80211_sdata_running(), some new code was introduced that uses netif_running() instead. Switch all these instances over. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: clean up ifdown/cleanup pathsJohannes Berg
There's a lot of redundant code in mac80211's interface cleanup/down, for example freeing AP beacons is done both when the interface is set DOWN as well as when it is torn down, of which only the former has any effect. Also, a bunch of things should be closer to where they matter, like the MLME timers that we should cancel when disassociating, rather than only when the interface is set DOWN. Clean up all this code. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: use subqueue helpersJohannes Berg
There are subqueue helpers so that we don't need to get the TX queue and then wake/stop it, use those helpers. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: allow changing port control protocolJohannes Berg
Some vendor specified mechanisms for 802.1X-style functionality use a different protocol than EAP (even if EAP is vendor-extensible). Support this in mac80211 via the cfg80211 API for it. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27cfg80211: allow changing port control protocolJohannes Berg
Some vendor specified mechanisms for 802.1X-style functionality use a different protocol than EAP (even if EAP is vendor-extensible). Allow setting the ethertype for the protocol when a driver has support for this. The default if unspecified is EAP, of course. Note: This is suitable only for station mode, not for AP implementation. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: allow vendor specific cipher suitesJohannes Berg
Allow drivers to specify their own set of cipher suites to advertise vendor-specific ciphers. The driver is then required to implement hardware crypto offload for it. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27cfg80211: allow vendor specific cipher suitesJohannes Berg
cfg80211 currently rejects all cipher suites it doesn't know about for key length checking purposes. This can lead to inconsistencies when a driver advertises an algorithm that cfg80211 doesn't know about. Remove this rejection so drivers can specify any algorithm they like. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: allow scan to complete from any contextJohannes Berg
The ieee80211_scan_completed() function was a frequent source of potential deadlocks, since it is called by drivers but may call back into drivers, so drivers had to make sure to call it without any locks held, which frequently lead to more complex code in drivers. Avoid that problem by allowing the function to be called in any context, and queueing the actual work it does. Also update the documentation for it to indicate this. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-27mac80211: remove unused scan expire defineJohannes Berg
Since cfg80211 manages the BSS list completely, this define hasn't been used for a long time and will never be used again. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-26gro: __napi_gro_receive() optimizationsEric Dumazet
compare_ether_header() can have a special implementation on 64 bit arches if CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS is defined. __napi_gro_receive() and vlan_gro_common() can avoid a conditional branch to perform device match. On x86_64, __napi_gro_receive() has now 38 instructions instead of 53 As gcc-4.4.3 still choose to not inline it, add inline keyword to this performance critical function. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> CC: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26net/caif/cfrfml.c: use asm/unaligned.hJeff Mahoney
caif does not build on ia64 starting with 2.6.32-rc1. Using asm/unaligned.h instead of linux/unaligned/le_byteshift.h fixes the issue. include/linux/unaligned/le_byteshift.h:40:50: error: redefinition of 'get_unaligned_le16' include/linux/unaligned/le_byteshift.h:45:50: error: redefinition of 'get_unaligned_le32' include/linux/unaligned/le_byteshift.h:50:50: error: redefinition of 'get_unaligned_le64' include/linux/unaligned/le_byteshift.h:55:51: error: redefinition of 'put_unaligned_le16' include/linux/unaligned/le_byteshift.h:60:51: error: redefinition of 'put_unaligned_le32' include/linux/unaligned/le_byteshift.h:65:51: error: redefinition of 'put_unaligned_le64' include/linux/unaligned/le_struct.h:31:51: note: previous definition of 'put_unaligned_le64' was here Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26ax25: missplaced sock_put(sk)Bernard Pidoux F6BVP
This patch moves a missplaced sock_put(sk) after bh_unlock_sock(sk) like in other parts of AX25 driver. Signed-off-by: Bernard Pidoux <f6bvp@free.fr> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26net: use scnprintf() to avoid potential buffer overflowChangli Gao
strlcpy() returns the total length of the string they tried to create, so we should not use its return value without any check. scnprintf() returns the number of characters written into @buf not including the trailing '\0', so use it instead here. Signed-off-by: Changli Gao <xiaosuo@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26net/sctp: Use pr_fmt and pr_<level>Joe Perches
Change SCTP_DEBUG_PRINTK and SCTP_DEBUG_PRINTK_IPADDR to use do { print } while (0) guards. Add SCTP_DEBUG_PRINTK_CONT to fix errors in log when lines were continued. Add #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt Add a missing newline in "Failed bind hash alloc" Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26l2tp: test for ethernet header in l2tp_eth_dev_recv()Eric Dumazet
close https://bugzilla.kernel.org/show_bug.cgi?id=16529 Before calling dev_forward_skb(), we should make sure skb head contains at least an ethernet header, even if length included in upper layer said so. Use pskb_may_pull() to make sure this ethernet header is present in skb head. Reported-by: Thomas Heil <heil@terminal-consulting.de> Reported-by: Ian Campbell <Ian.Campbell@eu.citrix.com> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26ipvs: switch to GFP_KERNEL allocationsSimon Horman
Switch from GFP_ATOMIC allocations to GFP_KERNEL ones in ip_vs_add_service() and ip_vs_new_dest(), as we hold a mutex and are allowed to sleep in this context. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26IPVS: convert __ip_vs_securetcp_lock to a spinlockSimon Horman
Also rename __ip_vs_securetcp_lock to ip_vs_securetcp_lock. Spinlock conversion was suggested by Eric Dumazet. Acked-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26IPVS: convert __ip_vs_sched_lock to a spinlockSimon Horman
Also rename __ip_vs_sched_lock to ip_vs_sched_lock. Acked-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-26IPVS: ICMPv6 checksum calculationSimon Horman
Cc: Xiaoyu Du <tingsrain@gmail.com> Signed-off-by: Julian Anastasov <ja@ssi.bg> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-25tcp: select(writefds) don't hang up when a peer close connectionKOSAKI Motohiro
This issue come from ruby language community. Below test program hang up when only run on Linux. % uname -mrsv Linux 2.6.26-2-486 #1 Sat Dec 26 08:37:39 UTC 2009 i686 % ruby -rsocket -ve ' BasicSocket.do_not_reverse_lookup = true serv = TCPServer.open("127.0.0.1", 0) s1 = TCPSocket.open("127.0.0.1", serv.addr[1]) s2 = serv.accept s2.close s1.write("a") rescue p $! s1.write("a") rescue p $! Thread.new { s1.write("a") }.join' ruby 1.9.3dev (2010-07-06 trunk 28554) [i686-linux] #<Errno::EPIPE: Broken pipe> [Hang Here] FreeBSD, Solaris, Mac doesn't. because Ruby's write() method call select() internally. and tcp_poll has a bug. SUS defined 'ready for writing' of select() as following. | A descriptor shall be considered ready for writing when a call to an output | function with O_NONBLOCK clear would not block, whether or not the function | would transfer data successfully. That said, EPIPE situation is clearly one of 'ready for writing'. We don't have read-side issue because tcp_poll() already has read side shutdown care. | if (sk->sk_shutdown & RCV_SHUTDOWN) | mask |= POLLIN | POLLRDNORM | POLLRDHUP; So, Let's insert same logic in write side. - reference url http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/31065 http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/31068 Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-25tcp: fix three tcp sysctls tuningEric Dumazet
As discovered by Anton Blanchard, current code to autotune tcp_death_row.sysctl_max_tw_buckets, sysctl_tcp_max_orphans and sysctl_max_syn_backlog makes little sense. The bigger a page is, the less tcp_max_orphans is : 4096 on a 512GB machine in Anton's case. (tcp_hashinfo.bhash_size * sizeof(struct inet_bind_hashbucket)) is much bigger if spinlock debugging is on. Its wrong to select bigger limits in this case (where kernel structures are also bigger) bhash_size max is 65536, and we get this value even for small machines. A better ground is to use size of ehash table, this also makes code shorter and more obvious. Based on a patch from Anton, and another from David. Reported-and-tested-by: Anton Blanchard <anton@samba.org> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-25bridge: avoid ethtool on non running interfacestephen hemminger
If bridge port is offline, don't call ethtool to query speed. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-25bridge: fix locking commentStephen Hemminger
The carrier check is not called from work queue in current code. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-25net/netfilter/ipvs: Eliminate memory leakJulia Lawall
__ip_vs_service_get and __ip_vs_svc_fwm_get increment a reference count, so that reference count should be decremented before leaving the function in an error case. A simplified version of the semantic match that finds this problem is: (http://coccinelle.lip6.fr/) // <smpl> @r exists@ local idexpression x; expression E; identifier f1; iterator I; @@ x = __ip_vs_service_get(...); <... when != x when != true (x == NULL || ...) when != if (...) { <+...x...+> } when != I (...) { <+...x...+> } ( x == NULL | x == E | x->f1 ) ...> * return ...; // </smpl> Signed-off-by: Julia Lawall <julia@diku.dk> Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-25Merge branch 'master' of ↵John W. Linville
git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next-2.6 into for-davem Conflicts: drivers/net/wireless/libertas/if_sdio.c
2010-08-25mac80211: fix mesh advertisementJohannes Berg
When a mac80211-based driver advertises mesh mode support, this will be advertised to userspace. However, if mac80211 was compiled without mesh support, then that won't actually be true. Fix this by removing the bit for mesh if mesh isn't compiled in. Since this synchronizes what we advertise to cfg80211 and actually support, it means we can now rely on cfg80211's interface type checks and need not check again in mac80211. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-25mac80211: fix rcu-unsafe pointer dereferenceChristian Lamparter
This patch fixes a potential crash (null-pointer de- reference) which was introduced in my previous patch: "mac80211: AMPDU rx reorder timeout timer" During a BA teardown, the pointer to the soon-to-be-gone tid_ampdu_rx element will be nullified. Therefore the release timer mechanism has to be careful not to accidentally access the item without any RCU protection. Signed-off-by: Christian Lamparter <chunkeey@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-25nl80211: fix missing nestingJohannes Berg
commit 95a6ccbb46c70cff376684c752831c014c87029d Author: Johannes Berg <johannes.berg@intel.com> Date: Thu Aug 12 15:38:38 2010 +0200 cfg80211/mac80211: extensible frame processing introduced a netlink bug that caused parsing errors in userspace because it forgot to close a nesting, which would advertise a nesting length of zero to userspace, which then completely threw off parsing and led to Illegal nla->nla_type == 0 being printed by libnl. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-25mac80211: cancel restart_work in ieee80211_unregister_hwChristian Lamparter
Unlike most other workqueue-tasks, the restart_work is not scheduled onto mac80211's private per-interface workqueue, but onto one of the system-wide workqueues. Therefore the mac80211-stack has to cancel any pending restarts, before destroying the shared device context and handing back the memory. Otherwise - under very unlucky circumstances - there could be a stale work- item left, because some other kernel component might have delayed the execution of ieee80211_restart_work for too long. Signed-off-by: Christian Lamparter <chunkeey@googlemail.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-25mac80211: fix warning for un-used parameterWey-Yi Guy
mesh_hdr only used when CONFIG_MAC80211_MESH is defined Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-25net/mac80211: Use wiphy_<level>Joe Perches
Standardize logging messages from printk(KERN_<level> "%s: " fmt , wiphy_name(foo), args); to wiphy_<level>(foo, fmt, args); Signed-off-by: Joe Perches <joe@perches.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-25tcp: Combat per-cpu skew in orphan tests.David S. Miller
As reported by Anton Blanchard when we use percpu_counter_read_positive() to make our orphan socket limit checks, the check can be off by up to num_cpus_online() * batch (which is 32 by default) which on a 128 cpu machine can be as large as the default orphan limit itself. Fix this by doing the full expensive sum check if the optimized check triggers. Reported-by: Anton Blanchard <anton@samba.org> Signed-off-by: David S. Miller <davem@davemloft.net> Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
2010-08-24tc: add meta match on receive hashStephen Hemminger
Trivial extension to existing meta data match rules to allow matching on skb receive hash value. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-24net: ip_append_data() optimEric Dumazet
Compiler is not smart enough to avoid a conditional branch. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-24cfg80211/mac80211: extensible frame processingJohannes Berg
Allow userspace to register for more than just action frames by giving the frame subtype, and make it possible to use this in various modes as well. With some tweaks and some added functionality this will, in the future, also be usable in AP mode and be able to replace the cooked monitor interface currently used in that case. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-24mac80211: remove unused don't-encrypt flagJohannes Berg
When MFP is disabled, action frames will not be encrypted since they are management frames and the only management frames that can then be encrypted are authentication frames. Therefore, setting the don't-encrypt flag on action frames is unnecessary. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-24cfg80211: mark ieee80211_hdrlen constJohannes Berg
This function analyses only its single, value-passed argument, and has no side effects. Thus it can be const, which makes mac80211 smaller, for example: text data bss dec hex filename 362518 16720 884 380122 5ccda mac80211.ko (before) 362358 16720 884 379962 5cc3a mac80211.ko (after) a 160 byte saving in text size, and an optimisation because the function won't be called as often. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: John W. Linville <linville@tuxdriver.com>
2010-08-23ethtool: allow non-netadmin to query settingsstephen hemminger
The SNMP daemon uses ethtool to determine the speed of network interfaces. This fails on Debian (and probably elsewhere) because for security SNMP daemon runs as non-root user (snmp). Note: A similar patch was rejected previously because of a concern about the possibility that on some hardware querying the ethtool settings requires access to the PHY and could slow the machine down. But the security risk of requiring SNMP daemon (and related services) to run as root far out weighs the risk of denial-of-service. Signed-off-by: Stephen Hemminger <shemminger@vyatta.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-23net: copy_rtnl_link_stats64() simplificationEric Dumazet
No need to use a temporary struct rtnl_link_stats64 variable, just copy the source to skb buffer. Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Reviewed-by: Ben Hutchings <bhutchings@solarflare.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-23net_sched: act_csum: coding style cleanupChangli Gao
Signed-off-by: Changli Gao <xiaosuo@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-23pkt_sched: Make act_csum depend upon INET.David S. Miller
It uses ip_send_check() and stuff like that. Reported-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-23bridge: netfilter: fix a memory leakChangli Gao
nf_bridge_alloc() always reset the skb->nf_bridge, so we should always put the old one. Signed-off-by: Changli Gao <xiaosuo@gmail.com> Signed-off-by: Bart De Schuymer <bdschuym@pandora.be> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-23dccp ccid-2: Replace broken RTT estimator with better algorithmGerrit Renker
The current CCID-2 RTT estimator code is in parts broken and lags behind the suggestions in RFC2988 of using scaled variants for SRTT/RTTVAR. That code is replaced by the present patch, which reuses the Linux TCP RTT estimator code. Further details: ---------------- 1. The minimum RTO of previously one second has been replaced with TCP's, since RFC4341, sec. 5 says that the minimum of 1 sec. (suggested in RFC2988, 2.4) is not necessary. Instead, the TCP_RTO_MIN is used, which agrees with DCCP's concept of a default RTT (RFC 4340, 3.4). 2. The maximum RTO has been set to DCCP_RTO_MAX (64 sec), which agrees with RFC2988, (2.5). 3. De-inlined the function ccid2_new_ack(). 4. Added a FIXME: the RTT is sampled several times per Ack Vector, which will give the wrong estimate. It should be replaced with one sample per Ack. However, at the moment this can not be resolved easily, since - it depends on TX history code (which also needs some work), - the cleanest solution is not to use the `sent' time at all (saves 4 bytes per entry) and use DCCP timestamps / elapsed time to estimated the RTT, which however is non-trivial to get right (but needs to be done). Reasons for reusing the Linux TCP estimator algorithm: ------------------------------------------------------ Some time was spent to find a better alternative, using basic RFC2988 as a first step. Further analysis and experimentation showed that the Linux TCP RTO estimator is superior to a basic RFC2988 implementation. A summary is on http://www.erg.abdn.ac.uk/users/gerrit/dccp/notes/ccid2/rto_estimator/ In addition, this estimator fared well in a recent empirical evaluation: Rewaskar, Sushant, Jasleen Kaur and F. Donelson Smith. A Performance Study of Loss Detection/Recovery in Real-world TCP Implementations. Proceedings of 15th IEEE International Conference on Network Protocols (ICNP-07), 2007. Thus there is significant benefit in reusing the existing TCP code. Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2010-08-23dccp ccid-2: Simplify dec_pipe and rearming of RTO timerGerrit Renker
This removes the dec_pipe function and improves the way the RTO timer is rearmed when a new acknowledgment comes in. Details and justification for removal: -------------------------------------- 1) The BUG_ON in dec_pipe is never triggered: pipe is only decremented for TX history entries between tail and head, for which it had previously been incremented in tx_packet_sent; and it is not decremented twice for the same entry, since it is - either decremented when a corresponding Ack Vector cell in state 0 or 1 was received (and then ccid2s_acked==1), - or it is decremented when ccid2s_acked==0, as part of the loss detection in tx_packet_recv (and hence it can not have been decremented earlier). 2) Restarting the RTO timer happens for every single entry in each Ack Vector parsed by tx_packet_recv (according to RFC 4340, 11.4 this can happen up to 16192 times per Ack Vector). 3) The RTO timer should not be restarted when all outstanding data has been acknowledged. This is currently done similar to (2), in dec_pipe, when pipe has reached 0. The patch onsolidates the code which rearms the RTO timer, combining the segments from new_ack and dec_pipe. As a result, the code becomes clearer (compare with tcp_rearm_rto()). Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
generated by cgit (git 2.34.1) at 2025-09-14 04:27:19 +0000