The UDP early demux can leverate the rx dst cache even for
multicast unconnected sockets.
In such scenario the ipv4 source address is validated only on
the first packet in the given flow. After that, when we fetch
the dst entry from the socket rx cache, we stop enforcing
the rp_filter and we even start accepting any kind of martian
addresses.
Disabling the dst cache for unconnected multicast socket will
cause large performace regression, nearly reducing by half the
max ingress tput.
Instead we factor out a route helper to completely validate an
skb source address for multicast packets and we call it from
the UDP early demux for mcast packets landing on unconnected
sockets, after successful fetching the related cached dst entry.
This still gives a measurable, but limited performance
regression:
rp_filter = 0 rp_filter = 1
edmux disabled: 1182 Kpps 1127 Kpps
edmux before: 2238 Kpps 2238 Kpps
edmux after: 2037 Kpps 2019 Kpps
The above figures are on top of current net tree.
Applying the net-next commit 6e617de84e ("net: avoid a full
fib lookup when rp_filter is disabled.") the delta with
rp_filter == 0 will decrease even more.
Fixes: 421b3885bf ("udp: ipv4: Add udp early demux")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently no error is emitted, but this infrastructure will
used by the next patch to allow source address validation
for mcast sockets.
Since early demux can do a route lookup and an ipv4 route
lookup can return an error code this is consistent with the
current ipv4 route infrastructure.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When CONFIG_KASAN is enabled, the "--param asan-stack=1" causes rather large
stack frames in some functions. This goes unnoticed normally because
CONFIG_FRAME_WARN is disabled with CONFIG_KASAN by default as of commit
3f181b4d86 ("lib/Kconfig.debug: disable -Wframe-larger-than warnings with
KASAN=y").
The kernelci.org build bot however has the warning enabled and that led
me to investigate it a little further, as every build produces these warnings:
net/wireless/nl80211.c:4389:1: warning: the frame size of 2240 bytes is larger than 2048 bytes [-Wframe-larger-than=]
net/wireless/nl80211.c:1895:1: warning: the frame size of 3776 bytes is larger than 2048 bytes [-Wframe-larger-than=]
net/wireless/nl80211.c:1410:1: warning: the frame size of 2208 bytes is larger than 2048 bytes [-Wframe-larger-than=]
net/bridge/br_netlink.c:1282:1: warning: the frame size of 2544 bytes is larger than 2048 bytes [-Wframe-larger-than=]
Most of this problem is now solved in gcc-8, which can consolidate
the stack slots for the inline function arguments. On older compilers
we can add a workaround by declaring a local variable in each function
to pass the inline function argument.
Cc: stable@vger.kernel.org
Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81715
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
In linux-4.13, Wei worked hard to convert dst to a traditional
refcounted model, removing GC.
We now want to make sure a dst refcount can not transition from 0 back
to 1.
The problem here is that input path attached a not refcounted dst to an
skb. Then later, because packet is forwarded and hits skb_dst_force()
before exiting RCU section, we might try to take a refcount on one dst
that is about to be freed, if another cpu saw 1 -> 0 transition in
dst_release() and queued the dst for freeing after one RCU grace period.
Lets unify skb_dst_force() and skb_dst_force_safe(), since we should
always perform the complete check against dst refcount, and not assume
it is not zero.
Bugzilla : https://bugzilla.kernel.org/show_bug.cgi?id=197005
[ 989.919496] skb_dst_force+0x32/0x34
[ 989.919498] __dev_queue_xmit+0x1ad/0x482
[ 989.919501] ? eth_header+0x28/0xc6
[ 989.919502] dev_queue_xmit+0xb/0xd
[ 989.919504] neigh_connected_output+0x9b/0xb4
[ 989.919507] ip_finish_output2+0x234/0x294
[ 989.919509] ? ipt_do_table+0x369/0x388
[ 989.919510] ip_finish_output+0x12c/0x13f
[ 989.919512] ip_output+0x53/0x87
[ 989.919513] ip_forward_finish+0x53/0x5a
[ 989.919515] ip_forward+0x2cb/0x3e6
[ 989.919516] ? pskb_trim_rcsum.part.9+0x4b/0x4b
[ 989.919518] ip_rcv_finish+0x2e2/0x321
[ 989.919519] ip_rcv+0x26f/0x2eb
[ 989.919522] ? vlan_do_receive+0x4f/0x289
[ 989.919523] __netif_receive_skb_core+0x467/0x50b
[ 989.919526] ? tcp_gro_receive+0x239/0x239
[ 989.919529] ? inet_gro_receive+0x226/0x238
[ 989.919530] __netif_receive_skb+0x4d/0x5f
[ 989.919532] netif_receive_skb_internal+0x5c/0xaf
[ 989.919533] napi_gro_receive+0x45/0x81
[ 989.919536] ixgbe_poll+0xc8a/0xf09
[ 989.919539] ? kmem_cache_free_bulk+0x1b6/0x1f7
[ 989.919540] net_rx_action+0xf4/0x266
[ 989.919543] __do_softirq+0xa8/0x19d
[ 989.919545] irq_exit+0x5d/0x6b
[ 989.919546] do_IRQ+0x9c/0xb5
[ 989.919548] common_interrupt+0x93/0x93
[ 989.919548] </IRQ>
Similarly dst_clone() can use dst_hold() helper to have additional
debugging, as a follow up to commit 44ebe79149 ("net: add debug
atomic_inc_not_zero() in dst_hold()")
In net-next we will convert dst atomic_t to refcount_t for peace of
mind.
Fixes: a4c2fd7f78 ("net: remove DST_NOCACHE flag")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Wei Wang <weiwan@google.com>
Reported-by: Paweł Staszewski <pstaszewski@itcare.pl>
Bisected-by: Paweł Staszewski <pstaszewski@itcare.pl>
Acked-by: Wei Wang <weiwan@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
remove tcp_may_send_now and tcp_snd_test that are no longer used
Fixes: 840a3cbe89 ("tcp: remove forward retransmit feature")
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 86fdb3448c ("sctp: ensure ep is not destroyed before doing the
dump") tried to fix an use-after-free issue by checking !sctp_sk(sk)->ep
with holding sock and sock lock.
But Paolo noticed that endpoint could be destroyed in sctp_rcv without
sock lock protection. It means the use-after-free issue still could be
triggered when sctp_rcv put and destroy ep after sctp_sock_dump checks
!ep, although it's pretty hard to reproduce.
I could reproduce it by mdelay in sctp_rcv while msleep in sctp_close
and sctp_sock_dump long time.
This patch is to add another param cb_done to sctp_for_each_transport
and dump ep->assocs with holding tsp after jumping out of transport's
traversal in it to avoid this issue.
It can also improve sctp diag dump to make it run faster, as no need
to save sk into cb->args[5] and keep calling sctp_for_each_transport
any more.
This patch is also to use int * instead of int for the pos argument
in sctp_for_each_transport, which could make postion increment only
in sctp_for_each_transport and no need to keep changing cb->args[2]
in sctp_sock_filter and sctp_sock_dump any more.
Fixes: 86fdb3448c ("sctp: ensure ep is not destroyed before doing the dump")
Reported-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This code causes a static checker warning because Smatch doesn't trust
anything that comes from skb->data. I've reviewed this code and I do
think skb->data can be controlled by the user here.
The sctp_event_subscribe struct has 13 __u8 fields and we want to see
if ours is non-zero. sn_type can be any value in the 0-USHRT_MAX range.
We're subtracting SCTP_SN_TYPE_BASE which is 1 << 15 so we could read
either before the start of the struct or after the end.
This is a very old bug and it's surprising that it would go undetected
for so long but my theory is that it just doesn't have a big impact so
it would be hard to notice.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
gen estimator has been rewritten in commit 1c0d32fde5
("net_sched: gen_estimator: complete rewrite of rate estimators"),
the caller is no longer needed to wait for a grace period.
So this patch gets rid of it.
This also completely closes a race condition between action free
path and filter chain add/remove path for the following patch.
Because otherwise the nested RCU callback can't be caught by
rcu_barrier().
Please see also the comments in code.
Cc: Jiri Pirko <jiri@mellanox.com>
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter/IPVS fixes for net
The following patchset contains Netfilter/IPVS fixes for your net tree,
they are:
1) Fix SCTP connection setup when IPVS module is loaded and any scheduler
is registered, from Xin Long.
2) Don't create a SCTP connection from SCTP ABORT packets, also from
Xin Long.
3) WARN_ON() and drop packet, instead of BUG_ON() races when calling
nf_nat_setup_info(). This is specifically a longstanding problem
when br_netfilter with conntrack support is in place, patch from
Florian Westphal.
4) Avoid softlock splats via iptables-restore, also from Florian.
5) Revert NAT hashtable conversion to rhashtable, semantics of rhlist
are different from our simple NAT hashtable, this has been causing
problems in the recent Linux kernel releases. From Florian.
6) Add per-bucket spinlock for NAT hashtable, so at least we restore
one of the benefits we got from the previous rhashtable conversion.
7) Fix incorrect hashtable size in memory allocation in xt_hashlimit,
from Zhizhou Tian.
8) Fix build/link problems with hashlimit and 32-bit arches, to address
recent fallout from a new hashlimit mode, from Vishwanath Pai.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 870190a9ec.
It was not a good idea. The custom hash table was a much better
fit for this purpose.
A fast lookup is not essential, in fact for most cases there is no lookup
at all because original tuple is not taken and can be used as-is.
What needs to be fast is insertion and deletion.
rhlist removal however requires a rhlist walk.
We can have thousands of entries in such a list if source port/addresses
are reused for multiple flows, if this happens removal requests are so
expensive that deletions of a few thousand flows can take several
seconds(!).
The advantages that we got from rhashtable are:
1) table auto-sizing
2) multiple locks
1) would be nice to have, but it is not essential as we have at
most one lookup per new flow, so even a million flows in the bysource
table are not a problem compared to current deletion cost.
2) is easy to add to custom hash table.
I tried to add hlist_node to rhlist to speed up rhltable_remove but this
isn't doable without changing semantics. rhltable_remove_fast will
check that the to-be-deleted object is part of the table and that
requires a list walk that we want to avoid.
Furthermore, using hlist_node increases size of struct rhlist_head, which
in turn increases nf_conn size.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=196821
Reported-by: Ivan Babrou <ibobrik@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
* a remain-on-channel fix from Avi
* hwsim TX power fix from Beni
* null-PTR dereference with iTXQ in some rare configurations (Chunho)
* 40 MHz custom regdomain fixes (Emmanuel)
* look at right place in HT/VHT capability parsing (Igor)
* complete A-MPDU teardown properly (Ilan)
* Mesh ID Element ordering fix (Liad)
* avoid tracing warning in ht_dbg() (Sharon)
* fix print of assoc/reassoc (Simon)
* fix encrypted VLAN with iTXQ (myself)
* fix calling context of TX queue wake (myself)
* fix a deadlock with ath10k aggregation (myself)
-----BEGIN PGP SIGNATURE-----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=QYS4
-----END PGP SIGNATURE-----
Merge tag 'mac80211-for-davem-2017-09-07' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211
Johannes Berg says:
====================
Back from a long absence, so we have a number of things:
* a remain-on-channel fix from Avi
* hwsim TX power fix from Beni
* null-PTR dereference with iTXQ in some rare configurations (Chunho)
* 40 MHz custom regdomain fixes (Emmanuel)
* look at right place in HT/VHT capability parsing (Igor)
* complete A-MPDU teardown properly (Ilan)
* Mesh ID Element ordering fix (Liad)
* avoid tracing warning in ht_dbg() (Sharon)
* fix print of assoc/reassoc (Simon)
* fix encrypted VLAN with iTXQ (myself)
* fix calling context of TX queue wake (myself)
* fix a deadlock with ath10k aggregation (myself)
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Let switch drivers indicate how many TX queues they support. Some
switches, such as Broadcom Starfighter 2 are designed with 8 egress
queues. Future changes will allow us to leverage the queue mapping and
direct the transmission towards a particular queue.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
__skb_flow_dissect is riddled with gotos that make discerning the flow,
debugging, and extending the capability difficult. This patch
reorganizes things so that we only perform goto's after the two main
switch statements (no gotos within the cases now). It also eliminates
several goto labels so that there are only two labels that can be target
for goto.
Reported-by: Alexander Popov <alex.popov@linux.com>
Signed-off-by: Tom Herbert <tom@quantonium.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
We get a new link error in allmodconfig kernels after ftgmac100
started using the ncsi helpers:
ERROR: "ncsi_vlan_rx_kill_vid" [drivers/net/ethernet/faraday/ftgmac100.ko] undefined!
ERROR: "ncsi_vlan_rx_add_vid" [drivers/net/ethernet/faraday/ftgmac100.ko] undefined!
Related to that, we get another error when CONFIG_NET_NCSI is disabled:
drivers/net/ethernet/faraday/ftgmac100.c:1626:25: error: 'ncsi_vlan_rx_add_vid' undeclared here (not in a function); did you mean 'ncsi_start_dev'?
drivers/net/ethernet/faraday/ftgmac100.c:1627:26: error: 'ncsi_vlan_rx_kill_vid' undeclared here (not in a function); did you mean 'ncsi_vlan_rx_add_vid'?
This fixes both problems at once, using a 'static inline' stub helper
for the disabled case, and exporting the functions when they are present.
Fixes: 51564585d8 ("ftgmac100: Support NCSI VLAN filtering when available")
Fixes: 21acf63013 ("net/ncsi: Configure VLAN tag filter")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
With TXQs, the AP_VLAN interfaces are resolved to their owner AP
interface when enqueuing the frame, which makes sense since the
frame really goes out on that as far as the driver is concerned.
However, this introduces a problem: frames to be encrypted with
a VLAN-specific GTK will now be encrypted with the AP GTK, since
the information about which virtual interface to use to select
the key is taken from the TXQ.
Fix this by preserving info->control.vif and using that in the
dequeue function. This now requires doing the driver-mapping
in the dequeue as well.
Since there's no way to filter the frames that are sitting on a
TXQ, drop all frames, which may affect other interfaces, when an
AP_VLAN is removed.
Cc: stable@vger.kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Pablo Neira Ayuso says:
====================
Netfilter updates for next-net (part 2)
The following patchset contains Netfilter updates for net-next. This
patchset includes updates for nf_tables, removal of
CONFIG_NETFILTER_DEBUG and a new mode for xt_hashlimit. More
specifically, they:
1) Add new rate match mode for hashlimit, this introduces a new revision
for this match. The idea is to stop matching packets until ratelimit
criteria stands true. Patch from Vishwanath Pai.
2) Add ->select_ops indirection to nf_tables named objects, so we can
choose between different flavours of the same object type, patch from
Pablo M. Bermudo.
3) Shorter function names in nft_limit, basically:
s/nft_limit_pkt_bytes/nft_limit_bytes, also from Pablo M. Bermudo.
4) Add new stateful limit named object type, this allows us to create
limit policies that you can identify via name, also from Pablo.
5) Remove unused hooknum parameter in conntrack ->packet indirection.
From Florian Westphal.
6) Patches to remove CONFIG_NETFILTER_DEBUG and macros such as
IP_NF_ASSERT and IP_NF_ASSERT. From Varsha Rao.
7) Add nf_tables_updchain() helper function and use it from
nf_tables_newchain() to make it more maintainable. Similarly,
add nf_tables_addchain() and use it too.
8) Add new netlink NLM_F_NONREC flag, this flag should only be used for
deletion requests, specifically, to support non-recursive deletion.
Based on what we discussed during NFWS'17 in Faro.
9) Use NLM_F_NONREC from table and sets in nf_tables.
10) Support for recursive chain deletion. Table and set deletion
commands come with an implicit content flush on deletion, while
chains do not. This patch addresses this inconsistency by adding
the code to perform recursive chain deletions. This also comes with
the bits to deal with the new NLM_F_NONREC netlink flag.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch removes CONFIG_NETFILTER_DEBUG and _ASSERT() macros as they
are no longer required. Replace _ASSERT() macros with WARN_ON().
Signed-off-by: Varsha Rao <rvarsha016@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
This patch adds support for overloading stateful objects operations
through the select_ops() callback, just as it is implemented for
expressions.
This change is needed for upcoming additions to the stateful objects
infrastructure.
Signed-off-by: Pablo M. Bermudo Garay <pablombg@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Johan Hedberg says:
====================
pull request: bluetooth-next 2017-09-03
Here's one last bluetooth-next pull request for the 4.14 kernel:
- NULL pointer fix in ca8210 802.15.4 driver
- A few "const" fixes
- New Kconfig option for disabling legacy interfaces
Please let me know if there are any issues pulling. Thanks.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Pablo Neira Ayuso says:
====================
Netfilter updates for net-next
The following patchset contains Netfilter updates for your net-next
tree. Basically, updates to the conntrack core, enhancements for
nf_tables, conversion of netfilter hooks from linked list to array to
improve memory locality and asorted improvements for the Netfilter
codebase. More specifically, they are:
1) Add expection to hashes after timer initialization to prevent
access from another CPU that walks on the hashes and calls
del_timer(), from Florian Westphal.
2) Don't update nf_tables chain counters from hot path, this is only
used by the x_tables compatibility layer.
3) Get rid of nested rcu_read_lock() calls from netfilter hook path.
Hooks are always guaranteed to run from rcu read side, so remove
nested rcu_read_lock() where possible. Patch from Taehee Yoo.
4) nf_tables new ruleset generation notifications include PID and name
of the process that has updated the ruleset, from Phil Sutter.
5) Use skb_header_pointer() from nft_fib, so we can reuse this code from
the nf_family netdev family. Patch from Pablo M. Bermudo.
6) Add support for nft_fib in nf_tables netdev family, also from Pablo.
7) Use deferrable workqueue for conntrack garbage collection, to reduce
power consumption, from Patch from Subash Abhinov Kasiviswanathan.
8) Add nf_ct_expect_iterate_net() helper and use it. From Florian
Westphal.
9) Call nf_ct_unconfirmed_destroy only from cttimeout, from Florian.
10) Drop references on conntrack removal path when skbuffs has escaped via
nfqueue, from Florian.
11) Don't queue packets to nfqueue with dying conntrack, from Florian.
12) Constify nf_hook_ops structure, from Florian.
13) Remove neededlessly branch in nf_tables trace code, from Phil Sutter.
14) Add nla_strdup(), from Phil Sutter.
15) Rise nf_tables objects name size up to 255 chars, people want to use
DNS names, so increase this according to what RFC 1035 specifies.
Patch series from Phil Sutter.
16) Kill nf_conntrack_default_on, it's broken. Default on conntrack hook
registration on demand, suggested by Eric Dumazet, patch from Florian.
17) Remove unused variables in compat_copy_entry_from_user both in
ip_tables and arp_tables code. Patch from Taehee Yoo.
18) Constify struct nf_conntrack_l4proto, from Julia Lawall.
19) Constify nf_loginfo structure, also from Julia.
20) Use a single rb root in connlimit, from Taehee Yoo.
21) Remove unused netfilter_queue_init() prototype, from Taehee Yoo.
22) Use audit_log() instead of open-coding it, from Geliang Tang.
23) Allow to mangle tcp options via nft_exthdr, from Florian.
24) Allow to fetch TCP MSS from nft_rt, from Florian. This includes
a fix for a miscalculation of the minimal length.
25) Simplify branch logic in h323 helper, from Nick Desaulniers.
26) Calculate netlink attribute size for conntrack tuple at compile
time, from Florian.
27) Remove protocol name field from nf_conntrack_{l3,l4}proto structure.
From Florian.
28) Remove holes in nf_conntrack_l4proto structure, so it becomes
smaller. From Florian.
29) Get rid of print_tuple() indirection for /proc conntrack listing.
Place all the code in net/netfilter/nf_conntrack_standalone.c.
Patch from Florian.
30) Do not built in print_conntrack() if CONFIG_NF_CONNTRACK_PROCFS is
off. From Florian.
31) Constify most nf_conntrack_{l3,l4}proto helper functions, from
Florian.
32) Fix broken indentation in ebtables extensions, from Colin Ian King.
33) Fix several harmless sparse warning, from Florian.
34) Convert netfilter hook infrastructure to use array for better memory
locality, joint work done by Florian and Aaron Conole. Moreover, add
some instrumentation to debug this.
35) Batch nf_unregister_net_hooks() calls, to call synchronize_net once
per batch, from Florian.
36) Get rid of noisy logging in ICMPv6 conntrack helper, from Florian.
37) Get rid of obsolete NFDEBUG() instrumentation, from Varsha Rao.
38) Remove unused code in the generic protocol tracker, from Davide
Caratti.
I think I will have material for a second Netfilter batch in my queue if
time allow to make it fit in this merge window.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 1d6119baf0.
After reverting commit 6d7b857d54 ("net: use lib/percpu_counter API
for fragmentation mem accounting") then here is no need for this
fix-up patch. As percpu_counter is no longer used, it cannot
memory leak it any-longer.
Fixes: 6d7b857d54 ("net: use lib/percpu_counter API for fragmentation mem accounting")
Fixes: 1d6119baf0 ("net: fix percpu memory leaks")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 6d7b857d54.
There is a bug in fragmentation codes use of the percpu_counter API,
that can cause issues on systems with many CPUs.
The frag_mem_limit() just reads the global counter (fbc->count),
without considering other CPUs can have upto batch size (130K) that
haven't been subtracted yet. Due to the 3MBytes lower thresh limit,
this become dangerous at >=24 CPUs (3*1024*1024/130000=24).
The correct API usage would be to use __percpu_counter_compare() which
does the right thing, and takes into account the number of (online)
CPUs and batch size, to account for this and call __percpu_counter_sum()
when needed.
We choose to revert the use of the lib/percpu_counter API for frag
memory accounting for several reasons:
1) On systems with CPUs > 24, the heavier fully locked
__percpu_counter_sum() is always invoked, which will be more
expensive than the atomic_t that is reverted to.
Given systems with more than 24 CPUs are becoming common this doesn't
seem like a good option. To mitigate this, the batch size could be
decreased and thresh be increased.
2) The add_frag_mem_limit+sub_frag_mem_limit pairs happen on the RX
CPU, before SKBs are pushed into sockets on remote CPUs. Given
NICs can only hash on L2 part of the IP-header, the NIC-RXq's will
likely be limited. Thus, a fair chance that atomic add+dec happen
on the same CPU.
Revert note that commit 1d6119baf0 ("net: fix percpu memory leaks")
removed init_frag_mem_limit() and instead use inet_frags_init_net().
After this revert, inet_frags_uninit_net() becomes empty.
Fixes: 6d7b857d54 ("net: use lib/percpu_counter API for fragmentation mem accounting")
Fixes: 1d6119baf0 ("net: fix percpu memory leaks")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
After ip_route_input() calls ip_route_input_noref(), another
check on skb_dst() is done, but if this fails, we shouldn't
override the return code from ip_route_input_noref(), as it
could have been more specific (i.e. -EHOSTUNREACH).
This also saves one call to skb_dst_force_safe() and one to
skb_dst() in case the ip_route_input_noref() check fails.
Reported-by: Sabrina Dubroca <sdubroca@redhat.com>
Fixes: 9df16efadd ("ipv4: call dst_hold_safe() properly")
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Acked-by: Wei Wang <weiwan@google.com>
Acked-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
When a listener registers to the FIB notification chain it receives a
dump of the FIB entries and rules from existing address families by
invoking their dump operations.
While we call into these modules we need to make sure they aren't
removed. Do that by increasing their reference count before invoking
their dump operations and decrease it afterwards.
Fixes: 04b1d4e50e ("net: core: Make the FIB notification chain generic")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Steffen Klassert says:
====================
pull request (net-next): ipsec-next 2017-09-01
This should be the last ipsec-next pull request for this
release cycle:
1) Support netdevice ESP trailer removal when decryption
is offloaded. From Yossi Kuperman.
2) Fix overwritten return value of copy_sec_ctx().
Please pull or let me know if there are problems.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
This will be used by the IPv6 host table which will be introduced in the
following patches. The fields in the header are added per-use. This header
is global and can be reused by many drivers.
Signed-off-by: Arkadi Sharshevsky <arkadis@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TC filters when used as classifiers are bound to TC classes.
However, there is a hidden difference when adding them in different
orders:
1. If we add tc classes before its filters, everything is fine.
Logically, the classes exist before we specify their ID's in
filters, it is easy to bind them together, just as in the current
code base.
2. If we add tc filters before the tc classes they bind, we have to
do dynamic lookup in fast path. What's worse, this happens all
the time not just once, because on fast path tcf_result is passed
on stack, there is no way to propagate back to the one in tc filters.
This hidden difference hurts performance silently if we have many tc
classes in hierarchy.
This patch intends to close this gap by doing the reverse binding when
we create a new class, in this case we can actually search all the
filters in its parent, match and fixup by classid. And because
tcf_result is specific to each type of tc filter, we have to introduce
a new ops for each filter to tell how to bind the class.
Note, we still can NOT totally get rid of those class lookup in
->enqueue() because cgroup and flow filters have no way to determine
the classid at setup time, they still have to go through dynamic lookup.
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In conjunction with crypto offload [1], removing the ESP trailer by
hardware can potentially improve the performance by avoiding (1) a
cache miss incurred by reading the nexthdr field and (2) the necessity
to calculate the csum value of the trailer in order to keep skb->csum
valid.
This patch introduces the changes to the xfrm stack and merely serves
as an infrastructure. Subsequent patch to mlx5 driver will put this to
a good use.
[1] https://www.mail-archive.com/netdev@vger.kernel.org/msg175733.html
Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Typically, each TC filter has its own action. All the actions of the
same type are saved in its hash table. But the hash buckets are too
small that it degrades to a list. And the performance is greatly
affected. For example, it takes about 0m11.914s to insert 64K rules.
If we convert the hash table to IDR, it only takes about 0m1.500s.
The improvement is huge.
But please note that the test result is based on previous patch that
cls_flower uses IDR.
Signed-off-by: Chris Mi <chrism@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 45f119bf93.
Eric Dumazet says:
We found at Google a significant regression caused by
45f119bf93 tcp: remove header prediction
In typical RPC (TCP_RR), when a TCP socket receives data, we now call
tcp_ack() while we used to not call it.
This touches enough cache lines to cause a slowdown.
so problem does not seem to be HP removal itself but the tcp_ack()
call. Therefore, it might be possible to remove HP after all, provided
one finds a way to elide tcp_ack for most cases.
Reported-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
This change was a followup to the header prediction removal,
so first revert this as a prerequisite to back out hp removal.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Florian reported UDP xmit drops that could be root caused to the
too small neigh limit.
Current limit is 64 KB, meaning that even a single UDP socket would hit
it, since its default sk_sndbuf comes from net.core.wmem_default
(~212992 bytes on 64bit arches).
Once ARP/ND resolution is in progress, we should allow a little more
packets to be queued, at least for one producer.
Once neigh arp_queue is filled, a rogue socket should hit its sk_sndbuf
limit and either block in sendmsg() or return -EAGAIN.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
NSH (Network Service Header)[1] is a new protocol for service
function chaining, it can be handled as a L3 protocol like
IPv4 and IPv6, Eth + NSH + Inner packet or VxLAN-gpe + NSH +
Inner packet are two typical use cases.
This patch adds NSH header structures and helpers for NSH GSO
support and Open vSwitch NSH support.
[1] https://datatracker.ietf.org/doc/draft-ietf-sfc-nsh/
[Jiri: added nsh_hdr() helper and renamed the header struct to "struct
nshhdr" to match the usual pattern. Removed packet type defines, these are
now shared with VXLAN-GPE.]
Signed-off-by: Yi Yang <yi.y.yang@intel.com>
Signed-off-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The values are shared between VXLAN-GPE and NSH. Originally probably by
coincidence but I notified both working groups about this last year and they
seem to keep the values in sync since then.
Hopefully they'll get a single IANA registry for the values, too. (I asked
them for that.)
Factor out the code to be shared by the NSH implementation.
NSH and MPLS values are added in this patch, too. For MPLS, the drafts
incorrectly assign only a single value, while we have two MPLS ethertypes.
I raised the problem with both groups. For now, I assume the value is for
unicast.
Signed-off-by: Jiri Benc <jbenc@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow a client call that failed on network error to be retried, provided
that the Tx queue still holds DATA packet 1. This allows an operation to
be submitted to another server or another address for the same server
without having to repackage and re-encrypt the data so far processed.
Two new functions are provided:
(1) rxrpc_kernel_check_call() - This is used to find out the completion
state of a call to guess whether it can be retried and whether it
should be retried.
(2) rxrpc_kernel_retry_call() - Disconnect the call from its current
connection, reset the state and submit it as a new client call to a
new address. The new address need not match the previous address.
A call may be retried even if all the data hasn't been loaded into it yet;
a partially constructed will be retained at the same point it was at when
an error condition was detected. msg_data_left() can be used to find out
how much data was packaged before the error occurred.
Signed-off-by: David Howells <dhowells@redhat.com>
Add a callback to rxrpc_kernel_send_data() so that a kernel service can get
a notification that the AF_RXRPC call has transitioned out the Tx phase and
is now waiting for a reply or a final ACK.
This is called from AF_RXRPC with the call state lock held so the
notification is guaranteed to come before any reply is passed back.
Further, modify the AFS filesystem to make use of this so that we don't have
to change the afs_call state before sending the last bit of data.
Signed-off-by: David Howells <dhowells@redhat.com>
Make use of the ndo_vlan_rx_{add,kill}_vid callbacks to have the NCSI
stack process new VLAN tags and configure the channel VLAN filter
appropriately.
Several VLAN tags can be set and a "Set VLAN Filter" packet must be sent
for each one, meaning the ncsi_dev_state_config_svf state must be
repeated. An internal list of VLAN tags is maintained, and compared
against the current channel's ncsi_channel_filter in order to keep track
within the state. VLAN filters are removed in a similar manner, with the
introduction of the ncsi_dev_state_config_clear_vids state. The maximum
number of VLAN tag filters is determined by the "Get Capabilities"
response from the channel.
Signed-off-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
And finally, move the irda include files into
drivers/staging/irda/include/net/irda. Yes, it's a long path, but it
makes it easy for us to just add a Makefile directory path addition and
all of the net and drivers code "just works".
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit c5cff8561d adds rcu grace period before freeing fib6_node. This
generates a new sparse warning on rt->rt6i_node related code:
net/ipv6/route.c:1394:30: error: incompatible types in comparison
expression (different address spaces)
./include/net/ip6_fib.h:187:14: error: incompatible types in comparison
expression (different address spaces)
This commit adds "__rcu" tag for rt6i_node and makes sure corresponding
rcu API is used for it.
After this fix, sparse no longer generates the above warning.
Fixes: c5cff8561d ("ipv6: add rcu grace period before freeing fib6_node")
Signed-off-by: Wei Wang <weiwan@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Similar to gre, vxlan, geneve, ipip tunnels, allow ERSPAN tunnels to
operate in 'collect metadata' mode. bpf_skb_[gs]et_tunnel_key() helpers
can make use of it right away. OVS can use it as well in the future.
Signed-off-by: William Tu <u9012063@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This converts the storage and layout of netfilter hook entries from a
linked list to an array. After this commit, hook entries will be
stored adjacent in memory. The next pointer is no longer required.
The ops pointers are stored at the end of the array as they are only
used in the register/unregister path and in the legacy br_netfilter code.
nf_unregister_net_hooks() is slower than needed as it just calls
nf_unregister_net_hook in a loop (i.e. at least n synchronize_net()
calls), this will be addressed in followup patch.
Test setup:
- ixgbe 10gbit
- netperf UDP_STREAM, 64 byte packets
- 5 hooks: (raw + mangle prerouting, mangle+filter input, inet filter):
empty mangle and raw prerouting, mangle and filter input hooks:
353.9
this patch:
364.2
Signed-off-by: Aaron Conole <aconole@bytheb.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Currently, in the udp6 code, the dst cookie is not initialized/updated
concurrently with the RX dst used by early demux.
As a result, the dst_check() in the early_demux path always fails,
the rx dst cache is always invalidated, and we can't really
leverage significant gain from the demux lookup.
Fix it adding udp6 specific variant of sk_rx_dst_set() and use it
to set the dst cookie when the dst entry is really changed.
The issue is there since the introduction of early demux for ipv6.
Fixes: 5425077d73 ("net: ipv6: Add early demux handler for UDP unicast")
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It is ugly to hide a u32-filter-specific pointer inside Qdisc,
this breaks the TC layers:
1. Qdisc is a generic representation, should not have any specific
data of any type
2. Qdisc layer is above filter layer, should only save filters in
the list of struct tcf_proto.
This pointer is used as the head of the chain of u32 hash tables,
that is struct tc_u_hnode, because u32 filter is very special,
it allows to create multiple hash tables within one qdisc and
across multiple u32 filters.
Instead of using this ugly pointer, we can just save it in a global
hash table key'ed by (dev ifindex, qdisc handle), therefore we can
still treat it as a per qdisc basis data structure conceptually.
Of course, because of network namespaces, this key is not unique
at all, but it is fine as we already have a pointer to Qdisc in
struct tc_u_common, we can just compare the pointers when collision.
And this only affects slow paths, has no impact to fast path,
thanks to the pointer ->tp_c.
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For TC classes, their ->get() and ->put() are always paired, and the
reference counting is completely useless, because:
1) For class modification and dumping paths, we already hold RTNL lock,
so all of these ->get(),->change(),->put() are atomic.
2) For filter bindiing/unbinding, we use other reference counter than
this one, and they should have RTNL lock too.
3) For ->qlen_notify(), it is special because it is called on ->enqueue()
path, but we already hold qdisc tree lock there, and we hold this
tree lock when graft or delete the class too, so it should not be gone
or changed until we release the tree lock.
Therefore, this patch removes ->get() and ->put(), but:
1) Adds a new ->find() to find the pointer to a class by classid, no
refcnt.
2) Move the original class destroy upon the last refcnt into ->delete(),
right after releasing tree lock. This is fine because the class is
already removed from hash when holding the lock.
For those who also use ->put() as ->unbind(), just rename them to reflect
this change.
Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Paolo Abeni