This patch enables fast retransmissions after one dupACK for
TCP if the stream is identified as thin. This will reduce
latencies for thin streams that are not able to trigger fast
retransmissions due to high packet interarrival time. This
mechanism is only active if enabled by iocontrol or syscontrol
and the stream is identified as thin.
Signed-off-by: Andreas Petlund <apetlund@simula.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch will make TCP use only linear timeouts if the
stream is thin. This will help to avoid the very high latencies
that thin stream suffer because of exponential backoff. This
mechanism is only active if enabled by iocontrol or syscontrol
and the stream is identified as thin. A maximum of 6 linear
timeouts is tried before exponential backoff is resumed.
Signed-off-by: Andreas Petlund <apetlund@simula.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
Make remaining netlink policies as const.
Fixup coding style where needed.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Dunno, what was the idea, it wasn't used for a long time.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ICMP6 MIB statistics was per-netns for quite a time.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The lock used in unix_state_lock() is a spin_lock not reader-writer.
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As reported by Randy Dunlap <randy.dunlap@oracle.com>, compilation
of nf_defrag_ipv4 fails with:
include/net/netfilter/nf_conntrack.h:94: error: field 'ct_general' has incomplete type
include/net/netfilter/nf_conntrack.h:178: error: 'const struct sk_buff' has no member named 'nfct'
include/net/netfilter/nf_conntrack.h:185: error: implicit declaration of function 'nf_conntrack_put'
include/net/netfilter/nf_conntrack.h:294: error: 'const struct sk_buff' has no member named 'nfct'
net/ipv4/netfilter/nf_defrag_ipv4.c:45: error: 'struct sk_buff' has no member named 'nfct'
net/ipv4/netfilter/nf_defrag_ipv4.c:46: error: 'struct sk_buff' has no member named 'nfct'
net/nf_conntrack.h must not be included with NF_CONNTRACK=n, add a
few #ifdefs. Long term the header file should be fixed to be usable
even with NF_CONNTRACK=n.
Tested-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Enhance IPVS to load balance SCTP transport protocol packets. This is done
based on the SCTP rfc 4960. All possible control chunks have been taken
care. The state machine used in this code looks some what lengthy. I tried
to make the state machine easy to understand.
Signed-off-by: Venkata Mohan Reddy Koppula <mohanreddykv@gmail.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Only used for writing, so convert to spinlock
Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Export sk_attach_filter/sk_detach_filter routines,
so that tun module can use them.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Traffic (tcp) doesnot start on a vlan interface when gro is enabled.
Even the tcp handshake was not taking place.
This is because, the eth_type_trans call before the netif_receive_skb
in napi_gro_finish() resets the skb->dev to napi->dev from the previously
set vlan netdev interface. This causes the ip_route_input to drop the
incoming packet considering it as a packet coming from a martian source.
I could repro this on 2.6.32.7 (stable) and 2.6.33-rc7.
With this fix, the traffic starts and the test runs fine on both vlan
and non-vlan interfaces.
CC: Herbert Xu <herbert@gondor.apana.org.au>
CC: Patrick McHardy <kaber@trash.net>
Signed-off-by: Ajit Khaparde <ajitk@serverengines.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
As reported by Alexey Dobriyan:
--------------------
setkey now takes several seconds to run this simple script
and it spits "recv: Resource temporarily unavailable" messages.
#!/usr/sbin/setkey -f
flush;
spdflush;
add A B ipcomp 44 -m tunnel -C deflate;
add B A ipcomp 45 -m tunnel -C deflate;
spdadd A B any -P in ipsec
ipcomp/tunnel/192.168.1.2-192.168.1.3/use;
spdadd B A any -P out ipsec
ipcomp/tunnel/192.168.1.3-192.168.1.2/use;
--------------------
Obviously applications want the events even when the table
is empty. So we cannot make this behavioral change.
Signed-off-by: David S. Miller <davem@davemloft.net>
The n-tuple list should be flushed if and only if the ETH_RESET_FILTER
flag is set and the driver is able to reset filtering/flow direction
hardware without also resetting a component whose flag is not set.
This test is best left to the driver.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 2eff25c18c
(netfilter: xt_hashlimit: fix race condition and simplify locking)
added a mutex deadlock :
htable_create() is called with hashlimit_mutex already locked
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The net being checked there is dev_net(dev) and thus this if
is always false.
Fits both net and net-next trees.
Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
call_rcu() will unconditionally reinitialize RCU head anyway.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add __percpu sparse annotations to net.
These annotations are to make sparse consider percpu variables to be
in a different address space and warn if accessed without going
through percpu accessors. This patch doesn't affect normal builds.
The macro and type tricks around snmp stats make things a bit
interesting. DEFINE/DECLARE_SNMP_STAT() macros mark the target field
as __percpu and SNMP_UPD_PO_STATS() macro is updated accordingly. All
snmp_mib_*() users which used to cast the argument to (void **) are
updated to cast it to (void __percpu **).
Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: David S. Miller <davem@davemloft.net>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Cc: Vlad Yasevich <vladislav.yasevich@hp.com>
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric's version fixed it for pfkey. This one is for xfrm user.
I thought about amortizing those two get_acqseq()s but it seems
reasonable to have two of these sequence spaces for the two different
interfaces.
cheers,
jamal
commit d5168d5addbc999c94aacda8f28a4a173756a72b
Author: Jamal Hadi Salim <hadi@cyberus.ca>
Date: Tue Feb 16 06:51:22 2010 -0500
xfrm: avoid spinlock in get_acqseq() used by xfrm user
This is in the same spirit as commit 28aecb9d77
by Eric Dumazet.
Use atomic_inc_return() in get_acqseq() to avoid taking a spinlock
Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (22 commits)
be2net: set proper value to version field in req hdr
xfrm: Fix xfrm_state_clone leak
ipcomp: Avoid duplicate calls to ipcomp_destroy
ethtool: allow non-admin user to read GRO settings.
ixgbe: fix WOL register setup for 82599
ixgbe: Fix - Do not allow Rx FC on 82598 at 1G due to errata
sfc: Fix SFE4002 initialisation
mac80211: fix handling of null-rate control in rate_control_get_rate
inet: Remove bogus IGMPv3 report handling
iwlwifi: fix AMSDU Rx after paged Rx patch
tcp: fix ICMP-RTO war
via-velocity: Fix races on shared interrupts
via-velocity: Take spinlock on set coalesce
via-velocity: Remove unused IRQ status parameter from rx_srv and tx_srv
rtl8187: Add new device ID
iwmc3200wifi: Test of wrong pointer after kzalloc in iwm_mlme_update_bss_table()
ath9k: Fix sequence numbers for PAE frames
mac80211: fix deferred hardware scan requests
iwlwifi: Fix to set correct ht configuration
mac80211: Fix probe request filtering in IBSS mode
...
Stop computing the number of neighbour table settings we have by
counting the number of binary sysctls. This behaviour was silly
and meant that we could not add another neighbour table setting
without also adding another binary sysctl.
Don't pass the binary sysctl path for neighour table entries
into neigh_sysctl_register. These parameters are no longer
used and so are just dead code.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Stop using the binary sysctl enumeartion in sysctl.h as an index into
a per interface array. This leads to unnecessary binary sysctl number
allocation, and a fragility in data structure and implementation
because of unnecessary coupling.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Same stuff as in ip_gre patch: receive hook can be called before netns
setup is done, oopsing in net_generic().
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
GRE protocol receive hook can be called right after protocol addition is done.
If netns stuff is not yet initialized, we're going to oops in
net_generic().
This is remotely oopsable if ip_gre is compiled as module and packet
comes at unfortunate moment of module loading.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
xfrm_state_clone calls kfree instead of xfrm_state_put to free
a failed state. Depending on the state of the failed state, it
can cause leaks to things like module references.
All states should be freed by xfrm_state_put past the point of
xfrm_init_state.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
When ipcomp_tunnel_attach fails we will call ipcomp_destroy twice.
This may lead to double-frees on certain structures.
As there is no reason to explicitly call ipcomp_destroy, this patch
removes it from ipcomp*.c and lets the standard xfrm_state destruction
take place.
This is based on the discovery and patch by Alexey Dobriyan.
Tested-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the required handlers to convert 32 bit
ebtables mark match and match target structs to 64bit layout.
Signed-off-by: Florian Westphal <fwestphal@astaro.com>
ebt_limit structure is larger on 64 bit systems due
to "long" type used in the (kernel-only) data section.
Setting .compatsize is enough in this case, these values
have no meaning in userspace.
Signed-off-by: Florian Westphal <fwestphal@astaro.com>
ebtables can be compiled to perform userspace-side padding of
structures. In that case, all the structures are already in the
'native' format expected by the kernel.
This tries to determine what format the userspace program is
using.
For most set/getsockopts, this can be done by checking
the len argument for sizeof(compat_ebt_replace) and
re-trying the native handler on error.
In case of EBT_SO_GET_ENTRIES, the native handler is tried first,
it will error out early when checking the *len argument
(the compat version has to defer this check until after
iterating over the kernel data set once, to adjust for all
the structure size differences).
As this would cause error printks, remove those as well, as
recommended by Bart de Schuymer.
Signed-off-by: Florian Westphal <fw@strlen.de>
Main code for 32 bit userland ebtables binary with 64 bit kernels
support.
Tested on x86_64 kernel only, using 64bit ebtables binary
for output comparision.
At least ebt_mark, m_mark and ebt_limit need CONFIG_COMPAT hooks, too.
remaining problem:
The ebtables userland makefile has:
ifeq ($(shell uname -m),sparc64)
CFLAGS+=-DEBT_MIN_ALIGN=8 -DKERNEL_64_USERSPACE_32
endif
struct ebt_replace, ebt_entry_match etc. then contain userland-side
padding, i.e. even if we are called from a 32 bit userland, the
structures may already be in the right format.
This problem is addressed in a follow-up patch.
Signed-off-by: Florian Westphal <fwestphal@astaro.com>
allows to call do_update_counters() from upcoming CONFIG_COMPAT
code instead of copy&pasting the same code.
Signed-off-by: Florian Westphal <fw@strlen.de>
once CONFIG_COMPAT support is added to ebtables, the new
copy_counters_to_user function can be called instead of duplicating
code.
Also remove last use of MEMPRINT, as requested by Bart De Schuymer.
Signed-off-by: Florian Westphal <fw@strlen.de>
once CONFIG_COMPAT support is merged this allows
to call do_replace_finish() after doing the CONFIG_COMPAT conversion
instead of copy & pasting this.
Signed-off-by: Florian Westphal <fw@strlen.de>
dev_ethtool() is currently using 604 bytes of stack, even with gcc-4.4.2
objdump -d vmlinux | scripts/checkstack.pl
...
0xc04bbc33 dev_ethtool [vmlinux]: 604
...
Adding noinline attributes to selected functions can reduce stack usage.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Addresses should be all digits.
Stops x25_bind using addresses containing characters.
Signed-off-by: Andrew Hendry <andrew.hendry@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
alloc_socket failures should return -ENOBUFS
a bad protocol should return -EINVAL
Signed-off-by: Andrew Hendry <andrew.hendry@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Observed similar behavior on SPD as previouly seen on SAD flushing..
This fixes it.
cheers,
jamal
commit 428b20432dc31bc2e01a94cd451cf5a2c00d2bf4
Author: Jamal Hadi Salim <hadi@cyberus.ca>
Date: Thu Feb 11 05:49:38 2010 -0500
xfrm: Flushing empty SPD generates false events
To see the effect make sure you have an empty SPD.
On window1 "ip xfrm mon" and on window2 issue "ip xfrm policy flush"
You get prompt back in window1 and you see the flush event on window2.
With this fix, you still get prompt on window1 but no event on window2.
Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
To see the effect make sure you have an empty SAD.
-On window1 "ip xfrm mon"
-on window2 issue "ip xfrm state flush"
You get prompt back in window1
and you see the flush event on window2.
With this fix, you still get prompt on window1 but no
event on window2.
I was tempted to return -ESRCH on window1 (which would
show "RTNETLINK answers: No such process") but didnt want
to change current behavior.
cheers,
jamal
commit 5f3dd4a772326166e1bcf54acc2391df00dc7ab5
Author: Jamal Hadi Salim <hadi@cyberus.ca>
Date: Thu Feb 11 04:41:36 2010 -0500
xfrm: Flushing empty SAD generates false events
To see the effect make sure you have an empty SAD.
On window1 "ip xfrm mon" and on window2 issue "ip xfrm state flush"
You get prompt back in window1 and you see the flush event on window2.
With this fix, you still get prompt on window1 but no event on window2.
Signed-off-by: Jamal Hadi Salim <hadi@cyberus.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
When no more memory can be allocated, fq_find() will return NULL and
increase the value of IPSTATS_MIB_REASMFAILS. In this case,
ipv6_frag_rcv() also increase the value of IPSTATS_MIB_REASMFAILS.
So, the patch deletes redundant counter of IPSTATS_MIB_REASMFAILS in fq_find().
and deletes the unused parameter of idev.
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
set_flags should check if the underlying device supports
n-tuple filter programming before setting the device flags
on the netdevice.
Signed-off-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>
We can allow a filter to be added successfully to the underlying
hardware, but still return an error if the cached list memory
allocation fails. This patch fixes that condition.
Signed-off-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>
with 32 bit userland and 64 bit kernels, it is unlikely but possible
that insertion of new rules fails even tough there are only about 2000
iptables rules.
This happens because the compat delta is using a short int.
Easily reproducible via "iptables -m limit" ; after about 2050
rules inserting new ones fails with -ELOOP.
Note that compat_delta included 2 bytes of padding on x86_64, so
structure size remains the same.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>
This will cause trouble once CONFIG_COMPAT support is added to ebtables.
xt_compat_*_offset() calculate the kernel/userland structure size delta
using:
XT_ALIGN(size) - COMPAT_XT_ALIGN(size)
If the match/target sizes are aligned at registration time,
delta is always zero.
Should have zero effect for existing systems: xtables uses
XT_ALIGN() whenever it deals with match/target sizes.
Signed-off-by: Florian Westphal <fwestphal@astaro.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
next_offset must be > 0, otherwise this loops forever.
The offset also contains the size of the ebt_entry structure
itself, so anything smaller is invalid.
Signed-off-by: Florian Westphal <fwestphal@astaro.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Andreas Petlund