Daniel Borkmann says:
====================
pull-request: bpf-next 2022-10-03
We've added 143 non-merge commits during the last 27 day(s) which contain
a total of 151 files changed, 8321 insertions(+), 1402 deletions(-).
The main changes are:
1) Add kfuncs for PKCS#7 signature verification from BPF programs, from Roberto Sassu.
2) Add support for struct-based arguments for trampoline based BPF programs,
from Yonghong Song.
3) Fix entry IP for kprobe-multi and trampoline probes under IBT enabled, from Jiri Olsa.
4) Batch of improvements to veristat selftest tool in particular to add CSV output,
a comparison mode for CSV outputs and filtering, from Andrii Nakryiko.
5) Add preparatory changes needed for the BPF core for upcoming BPF HID support,
from Benjamin Tissoires.
6) Support for direct writes to nf_conn's mark field from tc and XDP BPF program
types, from Daniel Xu.
7) Initial batch of documentation improvements for BPF insn set spec, from Dave Thaler.
8) Add a new BPF_MAP_TYPE_USER_RINGBUF map which provides single-user-space-producer /
single-kernel-consumer semantics for BPF ring buffer, from David Vernet.
9) Follow-up fixes to BPF allocator under RT to always use raw spinlock for the BPF
hashtab's bucket lock, from Hou Tao.
10) Allow creating an iterator that loops through only the resources of one
task/thread instead of all, from Kui-Feng Lee.
11) Add support for kptrs in the per-CPU arraymap, from Kumar Kartikeya Dwivedi.
12) Add a new kfunc helper for nf to set src/dst NAT IP/port in a newly allocated CT
entry which is not yet inserted, from Lorenzo Bianconi.
13) Remove invalid recursion check for struct_ops for TCP congestion control BPF
programs, from Martin KaFai Lau.
14) Fix W^X issue with BPF trampoline and BPF dispatcher, from Song Liu.
15) Fix percpu_counter leakage in BPF hashtab allocation error path, from Tetsuo Handa.
16) Various cleanups in BPF selftests to use preferred ASSERT_* macros, from Wang Yufen.
17) Add invocation for cgroup/connect{4,6} BPF programs for ICMP pings, from YiFei Zhu.
18) Lift blinding decision under bpf_jit_harden = 1 to bpf_capable(), from Yauheni Kaliuta.
19) Various libbpf fixes and cleanups including a libbpf NULL pointer deref, from Xin Liu.
* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (143 commits)
net: netfilter: move bpf_ct_set_nat_info kfunc in nf_nat_bpf.c
Documentation: bpf: Add implementation notes documentations to table of contents
bpf, docs: Delete misformatted table.
selftests/xsk: Fix double free
bpftool: Fix error message of strerror
libbpf: Fix overrun in netlink attribute iteration
selftests/bpf: Fix spelling mistake "unpriviledged" -> "unprivileged"
samples/bpf: Fix typo in xdp_router_ipv4 sample
bpftool: Remove unused struct event_ring_info
bpftool: Remove unused struct btf_attach_point
bpf, docs: Add TOC and fix formatting.
bpf, docs: Add Clang note about BPF_ALU
bpf, docs: Move Clang notes to a separate file
bpf, docs: Linux byteswap note
bpf, docs: Move legacy packet instructions to a separate file
selftests/bpf: Check -EBUSY for the recurred bpf_setsockopt(TCP_CONGESTION)
bpf: tcp: Stop bpf_setsockopt(TCP_CONGESTION) in init ops to recur itself
bpf: Refactor bpf_setsockopt(TCP_CONGESTION) handling into another function
bpf: Move the "cdg" tcp-cc check to the common sol_tcp_sockopt()
bpf: Add __bpf_prog_{enter,exit}_struct_ops for struct_ops trampoline
...
====================
Link: https://lore.kernel.org/r/20221003194915.11847-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When a bad bpf prog '.init' calls
bpf_setsockopt(TCP_CONGESTION, "itself"), it will trigger this loop:
.init => bpf_setsockopt(tcp_cc) => .init => bpf_setsockopt(tcp_cc) ...
... => .init => bpf_setsockopt(tcp_cc).
It was prevented by the prog->active counter before but the prog->active
detection cannot be used in struct_ops as explained in the earlier
patch of the set.
In this patch, the second bpf_setsockopt(tcp_cc) is not allowed
in order to break the loop. This is done by using a bit of
an existing 1 byte hole in tcp_sock to check if there is
on-going bpf_setsockopt(TCP_CONGESTION) in this tcp_sock.
Note that this essentially limits only the first '.init' can
call bpf_setsockopt(TCP_CONGESTION) to pick a fallback cc (eg. peer
does not support ECN) and the second '.init' cannot fallback to
another cc. This applies even the second
bpf_setsockopt(TCP_CONGESTION) will not cause a loop.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20220929070407.965581-5-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The more sockets we have in the hash table, the longer we spend looking
up the socket. While running a number of small workloads on the same
host, they penalise each other and cause performance degradation.
The root cause might be a single workload that consumes much more
resources than the others. It often happens on a cloud service where
different workloads share the same computing resource.
On EC2 c5.24xlarge instance (196 GiB memory and 524288 (1Mi / 2) ehash
entries), after running iperf3 in different netns, creating 24Mi sockets
without data transfer in the root netns causes about 10% performance
regression for the iperf3's connection.
thash_entries sockets length Gbps
524288 1 1 50.7
24Mi 48 45.1
It is basically related to the length of the list of each hash bucket.
For testing purposes to see how performance drops along the length,
I set 131072 (1Mi / 8) to thash_entries, and here's the result.
thash_entries sockets length Gbps
131072 1 1 50.7
1Mi 8 49.9
2Mi 16 48.9
4Mi 32 47.3
8Mi 64 44.6
16Mi 128 40.6
24Mi 192 36.3
32Mi 256 32.5
40Mi 320 27.0
48Mi 384 25.0
To resolve the socket lookup degradation, we introduce an optional
per-netns hash table for TCP, but it's just ehash, and we still share
the global bhash, bhash2 and lhash2.
With a smaller ehash, we can look up non-listener sockets faster and
isolate such noisy neighbours. In addition, we can reduce lock contention.
We can control the ehash size by a new sysctl knob. However, depending
on workloads, it will require very sensitive tuning, so we disable the
feature by default (net.ipv4.tcp_child_ehash_entries == 0). Moreover,
we can fall back to using the global ehash in case we fail to allocate
enough memory for a new ehash. The maximum size is 16Mi, which is large
enough that even if we have 48Mi sockets, the average list length is 3,
and regression would be less than 1%.
We can check the current ehash size by another read-only sysctl knob,
net.ipv4.tcp_ehash_entries. A negative value means the netns shares
the global ehash (per-netns ehash is disabled or failed to allocate
memory).
# dmesg | cut -d ' ' -f 5- | grep "established hash"
TCP established hash table entries: 524288 (order: 10, 4194304 bytes, vmalloc hugepage)
# sysctl net.ipv4.tcp_ehash_entries
net.ipv4.tcp_ehash_entries = 524288 # can be changed by thash_entries
# sysctl net.ipv4.tcp_child_ehash_entries
net.ipv4.tcp_child_ehash_entries = 0 # disabled by default
# ip netns add test1
# ip netns exec test1 sysctl net.ipv4.tcp_ehash_entries
net.ipv4.tcp_ehash_entries = -524288 # share the global ehash
# sysctl -w net.ipv4.tcp_child_ehash_entries=100
net.ipv4.tcp_child_ehash_entries = 100
# ip netns add test2
# ip netns exec test2 sysctl net.ipv4.tcp_ehash_entries
net.ipv4.tcp_ehash_entries = 128 # own a per-netns ehash with 2^n buckets
When more than two processes in the same netns create per-netns ehash
concurrently with different sizes, we need to guarantee the size in
one of the following ways:
1) Share the global ehash and create per-netns ehash
First, unshare() with tcp_child_ehash_entries==0. It creates dedicated
netns sysctl knobs where we can safely change tcp_child_ehash_entries
and clone()/unshare() to create a per-netns ehash.
2) Control write on sysctl by BPF
We can use BPF_PROG_TYPE_CGROUP_SYSCTL to allow/deny read/write on
sysctl knobs.
Note that the global ehash allocated at the boot time is spread over
available NUMA nodes, but inet_pernet_hashinfo_alloc() will allocate
pages for each per-netns ehash depending on the current process's NUMA
policy. By default, the allocation is done in the local node only, so
the per-netns hash table could fully reside on a random node. Thus,
depending on the NUMA policy the netns is created with and the CPU the
current thread is running on, we could see some performance differences
for highly optimised networking applications.
Note also that the default values of two sysctl knobs depend on the ehash
size and should be tuned carefully:
tcp_max_tw_buckets : tcp_child_ehash_entries / 2
tcp_max_syn_backlog : max(128, tcp_child_ehash_entries / 128)
As a bonus, we can dismantle netns faster. Currently, while destroying
netns, we call inet_twsk_purge(), which walks through the global ehash.
It can be potentially big because it can have many sockets other than
TIME_WAIT in all netns. Splitting ehash changes that situation, where
it's only necessary for inet_twsk_purge() to clean up TIME_WAIT sockets
in each netns.
With regard to this, we do not free the per-netns ehash in inet_twsk_kill()
to avoid UAF while iterating the per-netns ehash in inet_twsk_purge().
Instead, we do it in tcp_sk_exit_batch() after calling tcp_twsk_purge() to
keep it protocol-family-independent.
In the future, we could optimise ehash lookup/iteration further by removing
netns comparison for the per-netns ehash.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
While destroying netns, we call inet_twsk_purge() in tcp_sk_exit_batch()
and tcpv6_net_exit_batch() for AF_INET and AF_INET6. These commands
trigger the kernel to walk through the potentially big ehash twice even
though the netns has no TIME_WAIT sockets.
# ip netns add test
# ip netns del test
or
# unshare -n /bin/true >/dev/null
When tw_refcount is 1, we need not call inet_twsk_purge() at least
for the net. We can save such unneeded iterations if all netns in
net_exit_list have no TIME_WAIT sockets. This change eliminates
the tax by the additional unshare() described in the next patch to
guarantee the per-netns ehash size.
Tested:
# mount -t debugfs none /sys/kernel/debug/
# echo cleanup_net > /sys/kernel/debug/tracing/set_ftrace_filter
# echo inet_twsk_purge >> /sys/kernel/debug/tracing/set_ftrace_filter
# echo function > /sys/kernel/debug/tracing/current_tracer
# cat ./add_del_unshare.sh
for i in `seq 1 40`
do
(for j in `seq 1 100` ; do unshare -n /bin/true >/dev/null ; done) &
done
wait;
# ./add_del_unshare.sh
Before the patch:
# cat /sys/kernel/debug/tracing/trace_pipe
kworker/u128:0-8 [031] ...1. 174.162765: cleanup_net <-process_one_work
kworker/u128:0-8 [031] ...1. 174.240796: inet_twsk_purge <-cleanup_net
kworker/u128:0-8 [032] ...1. 174.244759: inet_twsk_purge <-tcp_sk_exit_batch
kworker/u128:0-8 [034] ...1. 174.290861: cleanup_net <-process_one_work
kworker/u128:0-8 [039] ...1. 175.245027: inet_twsk_purge <-cleanup_net
kworker/u128:0-8 [046] ...1. 175.290541: inet_twsk_purge <-tcp_sk_exit_batch
kworker/u128:0-8 [037] ...1. 175.321046: cleanup_net <-process_one_work
kworker/u128:0-8 [024] ...1. 175.941633: inet_twsk_purge <-cleanup_net
kworker/u128:0-8 [025] ...1. 176.242539: inet_twsk_purge <-tcp_sk_exit_batch
After:
# cat /sys/kernel/debug/tracing/trace_pipe
kworker/u128:0-8 [038] ...1. 428.116174: cleanup_net <-process_one_work
kworker/u128:0-8 [038] ...1. 428.262532: cleanup_net <-process_one_work
kworker/u128:0-8 [030] ...1. 429.292645: cleanup_net <-process_one_work
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We will soon introduce an optional per-netns ehash.
This means we cannot use tcp_hashinfo directly in most places.
Instead, access it via net->ipv4.tcp_death_row.hashinfo.
The access will be valid only while initialising tcp_hashinfo
itself and creating/destroying each netns.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We will soon introduce an optional per-netns ehash and access hash
tables via net->ipv4.tcp_death_row->hashinfo instead of &tcp_hashinfo
in most places.
It could harm the fast path because dereferences of two fields in net
and tcp_death_row might incur two extra cache line misses. To save one
dereference, let's place tcp_death_row back in netns_ipv4 and fetch
hashinfo via net->ipv4.tcp_death_row"."hashinfo.
Note tcp_death_row was initially placed in netns_ipv4, and commit
fbb8295248 ("tcp: allocate tcp_death_row outside of struct netns_ipv4")
changed it to a pointer so that we can fire TIME_WAIT timers after freeing
net. However, we don't do so after commit 04c494e68a ("Revert "tcp/dccp:
get rid of inet_twsk_purge()""), so we need not define tcp_death_row as a
pointer.
Also, we move refcount_dec_and_test(&tw_refcount) from tcp_sk_exit() to
tcp_sk_exit_batch() as a debug check.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch adds no functional change and cleans up some functions
that the following patches touch around so that we make them tidy
and easy to review/revert. The changes are
- Keep reverse christmas tree order
- Remove unnecessary init of port in inet_csk_find_open_port()
- Use req_to_sk() once in reqsk_queue_unlink()
- Use sock_net(sk) once in tcp_time_wait() and tcp_v[46]_connect()
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
While reading sysctl_tcp_abort_on_overflow, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its reader.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While reading sysctl_tcp_rfc1337, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its reader.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In tcp_create_openreq_child we adjust tcp_header_len for md5 using the
remote address in newsk. But that address is still 0 in newsk at this
point, and it is only set later by the callers (tcp_v[46]_syn_recv_sock).
Use the address from the request socket instead.
Fixes: cfb6eeb4c8 ("[TCP]: MD5 Signature Option (RFC2385) support.")
Signed-off-by: Francesco Ruggeri <fruggeri@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20220421005026.686A45EC01F2@us226.sjc.aristanetworks.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When setting RTO through BPF program, some SYN ACK packets were unaffected
and continued to use TCP_TIMEOUT_INIT constant. This patch adds timeout
option to struct request_sock. Option is initialized with TCP_TIMEOUT_INIT
and is reassigned through BPF using tcp_timeout_init call. SYN ACK
retransmits now use newly added timeout option.
Signed-off-by: Akhmat Karakotov <hmukos@yandex-team.ru>
Acked-by: Martin KaFai Lau <kafai@fb.com>
v2:
- Add timeout option to struct request_sock. Do not call
tcp_timeout_init on every syn ack retransmit.
v3:
- Use unsigned long for min. Bound tcp_timeout_init to TCP_RTO_MAX.
v4:
- Refactor duplicate code by adding reqsk_timeout function.
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp_minisocks.c hasn't use any macro or function declared in mm.h, module.h,
slab.h, sysctl.h, workqueue.h, static_key.h and inet_common.h. Thus, these
files can be removed from tcp_minisocks.c safely without affecting the
compilation of the net module.
Signed-off-by: Mianhan Liu <liumh1@shanghaitech.edu.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit adds two stats for the socket migration feature to evaluate the
effectiveness: LINUX_MIB_TCPMIGRATEREQ(SUCCESS|FAILURE).
If the migration fails because of the own_req race in receiving ACK and
sending SYN+ACK paths, we do not increment the failure stat. Then another
CPU is responsible for the req.
Link: https://lore.kernel.org/bpf/CAK6E8=cgFKuGecTzSCSQ8z3YJ_163C0uwO9yRvfDSE7vOe9mJA@mail.gmail.com/
Suggested-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch also changes the code to call reuseport_migrate_sock() and
inet_reqsk_clone(), but unlike the other cases, we do not call
inet_reqsk_clone() right after reuseport_migrate_sock().
Currently, in the receive path for TCP_NEW_SYN_RECV sockets, its listener
has three kinds of refcnt:
(A) for listener itself
(B) carried by reuqest_sock
(C) sock_hold() in tcp_v[46]_rcv()
While processing the req, (A) may disappear by close(listener). Also, (B)
can disappear by accept(listener) once we put the req into the accept
queue. So, we have to hold another refcnt (C) for the listener to prevent
use-after-free.
For socket migration, we call reuseport_migrate_sock() to select a listener
with (A) and to increment the new listener's refcnt in tcp_v[46]_rcv().
This refcnt corresponds to (C) and is cleaned up later in tcp_v[46]_rcv().
Thus we have to take another refcnt (B) for the newly cloned request_sock.
In inet_csk_complete_hashdance(), we hold the count (B), clone the req, and
try to put the new req into the accept queue. By migrating req after
winning the "own_req" race, we can avoid such a worst situation:
CPU 1 looks up req1
CPU 2 looks up req1, unhashes it, then CPU 1 loses the race
CPU 3 looks up req2, unhashes it, then CPU 2 loses the race
...
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20210612123224.12525-8-kuniyu@amazon.co.jp
Currently tcp_check_req can be called with obsolete req socket for which big
socket have been already created (because of CPU race or early demux
assigning req socket to multiple packets in gro batch).
Commit e0f9759f53 ("tcp: try to keep packet if SYN_RCV race
is lost") added retry in case when tcp_check_req is called for PSH|ACK packet.
But if client sends RST+ACK immediatly after connection being
established (it is performing healthcheck, for example) retry does not
occur. In that case tcp_check_req tries to close req socket,
leaving big socket active.
Fixes: e0f9759f53 ("tcp: try to keep packet if SYN_RCV race is lost")
Signed-off-by: Alexander Ovechkin <ovov@yandex-team.ru>
Reported-by: Oleg Senin <olegsenin@yandex-team.ru>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Because TCP-level resets only affect the subflow, there is a MPTCP
option to indicate that the MPTCP-level connection should be closed
immediately without a mptcp-level fin exchange.
This is the 'MPTCP fast close option'. It can be carried on ack
segments or TCP resets. In the latter case, its needed to parse mptcp
options also for reset packets so that MPTCP can act accordingly.
Next patch will add receive side fastclose support in MPTCP.
Acked-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
MP_JOIN subflows must not land into the accept queue.
Currently tcp_check_req() calls an mptcp specific helper
to detect such scenario.
Such helper leverages the subflow context to check for
MP_JOIN subflows. We need to deal also with MP JOIN
failures, even when the subflow context is not available
due allocation failure.
A possible solution would be changing the syn_recv_sock()
signature to allow returning a more descriptive action/
error code and deal with that in tcp_check_req().
Since the above need is MPTCP specific, this patch instead
uses a TCP request socket hole to add a MPTCP specific flag.
Such flag is used by the MPTCP syn_recv_sock() to tell
tcp_check_req() how to deal with the request socket.
This change is a no-op for !MPTCP build, and makes the
MPTCP code simpler. It allows also the next patch to deal
correctly with MP JOIN failure.
v1 -> v2:
- be more conservative on drop_req initialization (Mat)
RFC -> v1:
- move the drop_req bit inside tcp_request_sock (Eric)
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Reviewed-by: Christoph Paasch <cpaasch@apple.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Process the MP_JOIN option in a SYN packet with the same flow
as MP_CAPABLE but when the third ACK is received add the
subflow to the MPTCP socket subflow list instead of adding it to
the TCP socket accept queue.
The subflow is added at the end of the subflow list so it will not
interfere with the existing subflows operation and no data is
expected to be transmitted on it.
Co-developed-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Florian Westphal <fw@strlen.de>
Co-developed-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Peter Krystad <peter.krystad@linux.intel.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sparse reports warning at tcp_child_process()
warning: context imbalance in tcp_child_process() - unexpected unlock
The root cause is the missing annotation at tcp_child_process()
Add the missing __releases(&((child)->sk_lock.slock)) annotation
Signed-off-by: Jules Irenge <jbi.octave@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Prepare for cloning listening sockets that have their protocol callbacks
overridden by sk_msg. Child sockets must not inherit parent callbacks that
access state stored in sk_user_data owned by the parent.
Restore the child socket protocol callbacks before it gets hashed and any
of the callbacks can get invoked.
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/bpf/20200218171023.844439-4-jakub@cloudflare.com
This patch makes "struct tcp_congestion_ops" to be the first user
of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops
in bpf.
The BPF implemented tcp_congestion_ops can be used like
regular kernel tcp-cc through sysctl and setsockopt. e.g.
[root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion
net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic
net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic
net.ipv4.tcp_congestion_control = bpf_cubic
There has been attempt to move the TCP CC to the user space
(e.g. CCP in TCP). The common arguments are faster turn around,
get away from long-tail kernel versions in production...etc,
which are legit points.
BPF has been the continuous effort to join both kernel and
userspace upsides together (e.g. XDP to gain the performance
advantage without bypassing the kernel). The recent BPF
advancements (in particular BTF-aware verifier, BPF trampoline,
BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc)
possible in BPF. It allows a faster turnaround for testing algorithm
in the production while leveraging the existing (and continue growing)
BPF feature/framework instead of building one specifically for
userspace TCP CC.
This patch allows write access to a few fields in tcp-sock
(in bpf_tcp_ca_btf_struct_access()).
The optional "get_info" is unsupported now. It can be added
later. One possible way is to output the info with a btf-id
to describe the content.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com
There are few places where we fetch tp->snd_nxt while
this field can change from IRQ or other cpu.
We need to add READ_ONCE() annotations, and also make
sure write sides use corresponding WRITE_ONCE() to avoid
store-tearing.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are few places where we fetch tp->write_seq while
this field can change from IRQ or other cpu.
We need to add READ_ONCE() annotations, and also make
sure write sides use corresponding WRITE_ONCE() to avoid
store-tearing.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are few places where we fetch tp->copied_seq while
this field can change from IRQ or other cpu.
We need to add READ_ONCE() annotations, and also make
sure write sides use corresponding WRITE_ONCE() to avoid
store-tearing.
Note that tcp_inq_hint() was already using READ_ONCE(tp->copied_seq)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There are few places where we fetch tp->rcv_nxt while
this field can change from IRQ or other cpu.
We need to add READ_ONCE() annotations, and also make
sure write sides use corresponding WRITE_ONCE() to avoid
store-tearing.
Note that tcp_inq_hint() was already using READ_ONCE(tp->rcv_nxt)
syzbot reported :
BUG: KCSAN: data-race in tcp_poll / tcp_queue_rcv
write to 0xffff888120425770 of 4 bytes by interrupt on cpu 0:
tcp_rcv_nxt_update net/ipv4/tcp_input.c:3365 [inline]
tcp_queue_rcv+0x180/0x380 net/ipv4/tcp_input.c:4638
tcp_rcv_established+0xbf1/0xf50 net/ipv4/tcp_input.c:5616
tcp_v4_do_rcv+0x381/0x4e0 net/ipv4/tcp_ipv4.c:1542
tcp_v4_rcv+0x1a03/0x1bf0 net/ipv4/tcp_ipv4.c:1923
ip_protocol_deliver_rcu+0x51/0x470 net/ipv4/ip_input.c:204
ip_local_deliver_finish+0x110/0x140 net/ipv4/ip_input.c:231
NF_HOOK include/linux/netfilter.h:305 [inline]
NF_HOOK include/linux/netfilter.h:299 [inline]
ip_local_deliver+0x133/0x210 net/ipv4/ip_input.c:252
dst_input include/net/dst.h:442 [inline]
ip_rcv_finish+0x121/0x160 net/ipv4/ip_input.c:413
NF_HOOK include/linux/netfilter.h:305 [inline]
NF_HOOK include/linux/netfilter.h:299 [inline]
ip_rcv+0x18f/0x1a0 net/ipv4/ip_input.c:523
__netif_receive_skb_one_core+0xa7/0xe0 net/core/dev.c:5004
__netif_receive_skb+0x37/0xf0 net/core/dev.c:5118
netif_receive_skb_internal+0x59/0x190 net/core/dev.c:5208
napi_skb_finish net/core/dev.c:5671 [inline]
napi_gro_receive+0x28f/0x330 net/core/dev.c:5704
receive_buf+0x284/0x30b0 drivers/net/virtio_net.c:1061
read to 0xffff888120425770 of 4 bytes by task 7254 on cpu 1:
tcp_stream_is_readable net/ipv4/tcp.c:480 [inline]
tcp_poll+0x204/0x6b0 net/ipv4/tcp.c:554
sock_poll+0xed/0x250 net/socket.c:1256
vfs_poll include/linux/poll.h:90 [inline]
ep_item_poll.isra.0+0x90/0x190 fs/eventpoll.c:892
ep_send_events_proc+0x113/0x5c0 fs/eventpoll.c:1749
ep_scan_ready_list.constprop.0+0x189/0x500 fs/eventpoll.c:704
ep_send_events fs/eventpoll.c:1793 [inline]
ep_poll+0xe3/0x900 fs/eventpoll.c:1930
do_epoll_wait+0x162/0x180 fs/eventpoll.c:2294
__do_sys_epoll_pwait fs/eventpoll.c:2325 [inline]
__se_sys_epoll_pwait fs/eventpoll.c:2311 [inline]
__x64_sys_epoll_pwait+0xcd/0x170 fs/eventpoll.c:2311
do_syscall_64+0xcf/0x2f0 arch/x86/entry/common.c:296
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 7254 Comm: syz-fuzzer Not tainted 5.3.0+ #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Both tcp_v4_err() and tcp_v6_err() do the following operations
while they do not own the socket lock :
fastopen = tp->fastopen_rsk;
snd_una = fastopen ? tcp_rsk(fastopen)->snt_isn : tp->snd_una;
The problem is that without appropriate barrier, the compiler
might reload tp->fastopen_rsk and trigger a NULL deref.
request sockets are protected by RCU, we can simply add
the missing annotations and barriers to solve the issue.
Fixes: 168a8f5805 ("tcp: TCP Fast Open Server - main code path")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ctl packets sent on behalf of TIME_WAIT sockets currently
have a zero skb->priority, which can cause various problems.
In this patch we :
- add a tw_priority field in struct inet_timewait_sock.
- populate it from sk->sk_priority when a TIME_WAIT is created.
- For IPv4, change ip_send_unicast_reply() and its two
callers to propagate tw_priority correctly.
ip_send_unicast_reply() no longer changes sk->sk_priority.
- For IPv6, make sure TIME_WAIT sockets pass their tw_priority
field to tcp_v6_send_response() and tcp_v6_send_ack().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adding delays to TCP flows is crucial for studying behavior
of TCP stacks, including congestion control modules.
Linux offers netem module, but it has unpractical constraints :
- Need root access to change qdisc
- Hard to setup on egress if combined with non trivial qdisc like FQ
- Single delay for all flows.
EDT (Earliest Departure Time) adoption in TCP stack allows us
to enable a per socket delay at a very small cost.
Networking tools can now establish thousands of flows, each of them
with a different delay, simulating real world conditions.
This requires FQ packet scheduler or a EDT-enabled NIC.
This patchs adds TCP_TX_DELAY socket option, to set a delay in
usec units.
unsigned int tx_delay = 10000; /* 10 msec */
setsockopt(fd, SOL_TCP, TCP_TX_DELAY, &tx_delay, sizeof(tx_delay));
Note that FQ packet scheduler limits might need some tweaking :
man tc-fq
PARAMETERS
limit
Hard limit on the real queue size. When this limit is
reached, new packets are dropped. If the value is lowered,
packets are dropped so that the new limit is met. Default
is 10000 packets.
flow_limit
Hard limit on the maximum number of packets queued per
flow. Default value is 100.
Use of TCP_TX_DELAY option will increase number of skbs in FQ qdisc,
so packets would be dropped if any of the previous limit is hit.
Use of a jump label makes this support runtime-free, for hosts
never using the option.
Also note that TSQ (TCP Small Queues) limits are slightly changed
with this patch : we need to account that skbs artificially delayed
wont stop us providind more skbs to feed the pipe (netem uses
skb_orphan_partial() for this purpose, but FQ can not use this trick)
Because of that, using big delays might very well trigger
old bugs in TSO auto defer logic and/or sndbuf limited detection.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case autoflowlabel is in action, skb_get_hash_flowi6()
derives a non zero skb->hash to the flowlabel.
If skb->hash is zero, a flow dissection is performed.
Since all TCP skbs sent from ESTABLISH state inherit their
skb->hash from sk->sk_txhash, we better keep a copy
of sk->sk_txhash into the TIME_WAIT socket.
After this patch, ACK or RST packets sent on behalf of
a TIME_WAIT socket have the flowlabel that was previously
used by the flow.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add SPDX license identifiers to all files which:
- Have no license information of any form
- Have EXPORT_.*_SYMBOL_GPL inside which was used in the
initial scan/conversion to ignore the file
These files fall under the project license, GPL v2 only. The resulting SPDX
license identifier is:
GPL-2.0-only
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Linux implements RFC6298 and use an initial congestion window
of 1 upon establishing the connection if the SYNACK packet is
retransmitted 2 or more times. In cellular networks SYNACK timeouts
are often spurious if the wireless radio was dormant or idle. Also
some network path is longer than the default SYNACK timeout. In
both cases falsely starting with a minimal cwnd are detrimental
to performance.
This patch avoids doing so when the final ACK's TCP timestamp
indicates the original SYNACK was delivered. It remembers the
original SYNACK timestamp when SYNACK timeout has occurred and
re-uses the function to detect spurious SYN timeout conveniently.
Note that a server may receives multiple SYNs from and immediately
retransmits SYNACKs without any SYNACK timeout. This often happens
on when the client SYNs have timed out due to wireless delay
above. In this case since the server will still use the default
initial congestion (e.g. 10) because tp->undo_marker is reset in
tcp_init_metrics(). This is an intentional design because packets
are not lost but delayed.
This patch only covers regular TCP passive open. Fast Open is
supported in the next patch.
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>
This might speedup tcp_twsk_destructor() a bit,
avoiding a cache line miss.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we make sure all listeners have these fields cleared, then a clone
will also inherit zero values.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we make sure all listeners have proper tp->rack value,
then a clone will also inherit proper initial value.
Note that fresh sockets init tp->rack from tcp_init_sock()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we make sure all listeners have app_limited set to ~0U,
then a clone will also inherit proper initial value.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we make sure all listeners have these fields cleared, then a clone
will also inherit zero values.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All listeners have this field cleared already, since tcp_disconnect()
clears it and newly created sockets have also a zero value here.
So a clone will inherit a zero value here.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Passive connections can inherit proper value by cloning,
if we make sure all listeners have the proper values there.
tcp_disconnect() was setting snd_cwnd to 2, which seems
quite obsolete since IW10 adoption.
Also remove an obsolete comment.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we make sure a listener always has its mdev_us
field set to TCP_TIMEOUT_INIT, we do not need to rewrite
this field after a new clone is created.
tcp_disconnect() is very seldom used in real applications.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
All listeners have this field cleared already, since tcp_disconnect()
clears it and newly created sockets have also a zero value here.
So a clone will inherit a zero value here.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
New sockets have this field cleared, and tcp_disconnect()
calls tcp_write_queue_purge() which among other things
also clear tp->packets_out
So a listener is guaranteed to have this field cleared.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
If we make sure a listener always has its icsk_rto
field set to TCP_TIMEOUT_INIT, we do not need to rewrite
this field after a new clone is created.
tcp_disconnect() is very seldom used in real applications.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
New sockets get the field set to TCP_INFINITE_SSTHRESH in tcp_init_sock()
In case a socket had this field changed and transitions to TCP_LISTEN
state, tcp_disconnect() also makes sure snd_ssthresh is set to
TCP_INFINITE_SSTHRESH.
So a listener has this field set to TCP_INFINITE_SSTHRESH already.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RFC 1337 says:
''Ignore RST segments in TIME-WAIT state.
If the 2 minute MSL is enforced, this fix avoids all three hazards.''
So with net.ipv4.tcp_rfc1337=1, expected behaviour is to have TIME-WAIT sk
expire rather than removing it instantly when a reset is received.
However, Linux will also re-start the TIME-WAIT timer.
This causes connect to fail when tying to re-use ports or very long
delays (until syn retry interval exceeds MSL).
packetdrill test case:
// Demonstrate bogus rearming of TIME-WAIT timer in rfc1337 mode.
`sysctl net.ipv4.tcp_rfc1337=1`
0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
0.000 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
0.000 bind(3, ..., ...) = 0
0.000 listen(3, 1) = 0
0.100 < S 0:0(0) win 29200 <mss 1460,nop,nop,sackOK,nop,wscale 7>
0.100 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 7>
0.200 < . 1:1(0) ack 1 win 257
0.200 accept(3, ..., ...) = 4
// Receive first segment
0.310 < P. 1:1001(1000) ack 1 win 46
// Send one ACK
0.310 > . 1:1(0) ack 1001
// read 1000 byte
0.310 read(4, ..., 1000) = 1000
// Application writes 100 bytes
0.350 write(4, ..., 100) = 100
0.350 > P. 1:101(100) ack 1001
// ACK
0.500 < . 1001:1001(0) ack 101 win 257
// close the connection
0.600 close(4) = 0
0.600 > F. 101:101(0) ack 1001 win 244
// Our side is in FIN_WAIT_1 & waits for ack to fin
0.7 < . 1001:1001(0) ack 102 win 244
// Our side is in FIN_WAIT_2 with no outstanding data.
0.8 < F. 1001:1001(0) ack 102 win 244
0.8 > . 102:102(0) ack 1002 win 244
// Our side is now in TIME_WAIT state, send ack for fin.
0.9 < F. 1002:1002(0) ack 102 win 244
0.9 > . 102:102(0) ack 1002 win 244
// Peer reopens with in-window SYN:
1.000 < S 1000:1000(0) win 9200 <mss 1460,nop,nop,sackOK,nop,wscale 7>
// Therefore, reply with ACK.
1.000 > . 102:102(0) ack 1002 win 244
// Peer sends RST for this ACK. Normally this RST results
// in tw socket removal, but rfc1337=1 setting prevents this.
1.100 < R 1002:1002(0) win 244
// second syn. Due to rfc1337=1 expect another pure ACK.
31.0 < S 1000:1000(0) win 9200 <mss 1460,nop,nop,sackOK,nop,wscale 7>
31.0 > . 102:102(0) ack 1002 win 244
// .. and another RST from peer.
31.1 < R 1002:1002(0) win 244
31.2 `echo no timer restart;ss -m -e -a -i -n -t -o state TIME-WAIT`
// third syn after one minute. Time-Wait socket should have expired by now.
63.0 < S 1000:1000(0) win 9200 <mss 1460,nop,nop,sackOK,nop,wscale 7>
// so we expect a syn-ack & 3whs to proceed from here on.
63.0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK,nop,wscale 7>
Without this patch, 'ss' shows restarts of tw timer and last packet is
thus just another pure ack, more than one minute later.
This restores the original code from commit 283fd6cf0be690a83
("Merge in ANK networking jumbo patch") in netdev-vger-cvs.git .
For some reason the else branch was removed/lost in 1f28b683339f7
("Merge in TCP/UDP optimizations and [..]") and timer restart became
unconditional.
Reported-by: Michal Tesar <mtesar@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Using get_seconds() for timestamps is deprecated since it can lead
to overflows on 32-bit systems. While the interface generally doesn't
overflow until year 2106, the specific implementation of the TCP PAWS
algorithm breaks in 2038 when the intermediate signed 32-bit timestamps
overflow.
A related problem is that the local timestamps in CLOCK_REALTIME form
lead to unexpected behavior when settimeofday is called to set the system
clock backwards or forwards by more than 24 days.
While the first problem could be solved by using an overflow-safe method
of comparing the timestamps, a nicer solution is to use a monotonic
clocksource with ktime_get_seconds() that simply doesn't overflow (at
least not until 136 years after boot) and that doesn't change during
settimeofday().
To make 32-bit and 64-bit architectures behave the same way here, and
also save a few bytes in the tcp_options_received structure, I'm changing
the type to a 32-bit integer, which is now safe on all architectures.
Finally, the ts_recent_stamp field also (confusingly) gets used to store
a jiffies value in tcp_synq_overflow()/tcp_synq_no_recent_overflow().
This is currently safe, but changing the type to 32-bit requires
some small changes there to keep it working.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Tested: 'git grep tw_timeout' comes up empty and it builds :-)
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This version has some suggestions by Eric Dumazet:
- Use a local variable for the mark in IPv6 instead of ctl_sk to avoid SMP
races.
- Use the more elegant "IP4_REPLY_MARK(net, skb->mark) ?: sk->sk_mark"
statement.
- Factorize code as sk_fullsock() check is not necessary.
Aidan McGurn from Openwave Mobility systems reported the following bug:
"Marked routing is broken on customer deployment. Its effects are large
increase in Uplink retransmissions caused by the client never receiving
the final ACK to their FINACK - this ACK misses the mark and routes out
of the incorrect route."
Currently marks are added to sk_buffs for replies when the "fwmark_reflect"
sysctl is enabled. But not for TW sockets that had sk->sk_mark set via
setsockopt(SO_MARK..).
Fix this in IPv4/v6 by adding tw->tw_mark for TIME_WAIT sockets. Copy the the
original sk->sk_mark in __inet_twsk_hashdance() to the new tw->tw_mark location.
Then progate this so that the skb gets sent with the correct mark. Do the same
for resets. Give the "fwmark_reflect" sysctl precedence over sk->sk_mark so that
netfilter rules are still honored.
Signed-off-by: Jon Maxwell <jmaxwell37@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Exchange messages with hardware to program the TLS session
CPL handlers for messages received from chip.
Signed-off-by: Atul Gupta <atul.gupta@chelsio.com>
Signed-off-by: Michael Werner <werner@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
배석진 reported that in some situations, packets for a given 5-tuple
end up being processed by different CPUS.
This involves RPS, and fragmentation.
배석진 is seeing packet drops when a SYN_RECV request socket is
moved into ESTABLISH state. Other states are protected by socket lock.
This is caused by a CPU losing the race, and simply not caring enough.
Since this seems to occur frequently, we can do better and perform
a second lookup.
Note that all needed memory barriers are already in the existing code,
thanks to the spin_lock()/spin_unlock() pair in inet_ehash_insert()
and reqsk_put(). The second lookup must find the new socket,
unless it has already been accepted and closed by another cpu.
Note that the fragmentation could be avoided in the first place by
use of a correct TCP MSS option in the SYN{ACK} packet, but this
does not mean we can not be more robust.
Many thanks to 배석진 for a very detailed analysis.
Reported-by: 배석진 <soukjin.bae@samsung.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
First, rename __inet_twsk_hashdance() to inet_twsk_hashdance()
Then, remove one inet_twsk_put() by setting tw_refcnt to 3 instead
of 4, but adding a fat warning that we do not have the right to access
tw anymore after inet_twsk_hashdance()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Maciej Żenczykowski reported some panics in tcp_twsk_destructor()
that might be caused by the following bug.
timewait timer is pinned to the cpu, because we want to transition
timwewait refcount from 0 to 4 in one go, once everything has been
initialized.
At the time commit ed2e923945 ("tcp/dccp: fix timewait races in timer
handling") was merged, TCP was always running from BH habdler.
After commit 5413d1babe ("net: do not block BH while processing
socket backlog") we definitely can run tcp_time_wait() from process
context.
We need to block BH in the critical section so that the pinned timer
has still its purpose.
This bug is more likely to happen under stress and when very small RTO
are used in datacenter flows.
Fixes: 5413d1babe ("net: do not block BH while processing socket backlog")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Maciej Żenczykowski <maze@google.com>
Acked-by: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace the reordering distance measurement in packet unit with
sequence based approach. Previously it trackes the number of "packets"
toward the forward ACK (i.e. highest sacked sequence)in a state
variable "fackets_out".
Precisely measuring reordering degree on packet distance has not much
benefit, as the degree constantly changes by factors like path, load,
and congestion window. It is also complicated and prone to arcane bugs.
This patch replaces with sequence-based approach that's much simpler.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Priyaranjan Jha <priyarjha@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
FACK loss detection has been disabled by default and the
successor RACK subsumed FACK and can handle reordering better.
This patch removes FACK to simplify TCP loss recovery.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Priyaranjan Jha <priyarjha@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently TCP RACK loss detection does not work well if packets are
being reordered beyond its static reordering window (min_rtt/4).Under
such reordering it may falsely trigger loss recoveries and reduce TCP
throughput significantly.
This patch improves that by increasing and reducing the reordering
window based on DSACK, which is now supported in major TCP implementations.
It makes RACK's reo_wnd adaptive based on DSACK and no. of recoveries.
- If DSACK is received, increment reo_wnd by min_rtt/4 (upper bounded
by srtt), since there is possibility that spurious retransmission was
due to reordering delay longer than reo_wnd.
- Persist the current reo_wnd value for TCP_RACK_RECOVERY_THRESH (16)
no. of successful recoveries (accounts for full DSACK-based loss
recovery undo). After that, reset it to default (min_rtt/4).
- At max, reo_wnd is incremented only once per rtt. So that the new
DSACK on which we are reacting, is due to the spurious retx (approx)
after the reo_wnd has been updated last time.
- reo_wnd is tracked in terms of steps (of min_rtt/4), rather than
absolute value to account for change in rtt.
In our internal testing, we observed significant increase in throughput,
in scenarios where reordering exceeds min_rtt/4 (previous static value).
Signed-off-by: Priyaranjan Jha <priyarjha@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The SMC protocol [1] relies on the use of a new TCP experimental
option [2, 3]. With this option, SMC capabilities are exchanged
between peers during the TCP three way handshake. This patch adds
support for this experimental option to TCP.
References:
[1] SMC-R Informational RFC: http://www.rfc-editor.org/info/rfc7609
[2] Shared Use of TCP Experimental Options RFC 6994:
https://tools.ietf.org/rfc/rfc6994.txt
[3] IANA ExID SMCR:
http://www.iana.org/assignments/tcp-parameters/tcp-parameters.xhtml#tcp-exids
Signed-off-by: Ursula Braun <ubraun@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use BUG_ON instead of if condition followed by BUG in tcp_time_wait.
This issue was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds a new queue (list) that tracks the sent but not yet
acked or SACKed skbs for a TCP connection. The list is chronologically
ordered by skb->skb_mstamp (the head is the oldest sent skb).
This list will be used to optimize TCP Rack recovery, which checks
an skb's timestamp to judge if it has been lost and needs to be
retransmitted. Since TCP write queue is ordered by sequence instead
of sent time, RACK has to scan over the write queue to catch all
eligible packets to detect lost retransmission, and iterates through
SACKed skbs repeatedly.
Special cares for rare events:
1. TCP repair fakes skb transmission so the send queue needs adjusted
2. SACK reneging would require re-inserting SACKed skbs into the
send queue. For now I believe it's not worth the complexity to
make RACK work perfectly on SACK reneging, so we do nothing here.
3. Fast Open: currently for non-TFO, send-queue correctly queues
the pure SYN packet. For TFO which queues a pure SYN and
then a data packet, send-queue only queues the data packet but
not the pure SYN due to the structure of TFO code. This is okay
because the SYN receiver would never respond with a SACK on a
missing SYN (i.e. SYN is never fast-retransmitted by SACK/RACK).
In order to not grow sk_buff, we use an union for the new list and
_skb_refdst/destructor fields. This is a bit complicated because
we need to make sure _skb_refdst and destructor are properly zeroed
before skb is cloned/copied at transmit, and before being freed.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.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>
Like prequeue, I am not sure this is overly useful nowadays.
If we receive a train of packets, GRO will aggregate them if the
headers are the same (HP predates GRO by several years) so we don't
get a per-packet benefit, only a per-aggregated-packet one.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
prequeue is a tcp receive optimization that moves part of rx processing
from bh to process context.
This only works if the socket being processed belongs to a process that
is blocked in recv on that socket.
In practice, this doesn't happen anymore that often because nowadays
servers tend to use an event driven (epoll) model.
Even normal client applications (web browsers) commonly use many tcp
connections in parallel.
This has measureable impact only in netperf (which uses plain recv and
thus allows prequeue use) from host to locally running vm (~4%), however,
there were no changes when using netperf between two physical hosts with
ixgbe interfaces.
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds suppport for setting the initial advertized window from
within a BPF_SOCK_OPS program. This can be used to support larger
initial cwnd values in environments where it is known to be safe.
Signed-off-by: Lawrence Brakmo <brakmo@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We want to move some TCP sysctls to net namespaces in the future.
tcp_window_scaling, tcp_sack and tcp_timestamps being fetched
from tcp_parse_options(), we need to pass an extra parameter.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP Timestamps option is defined in RFC 7323
Traditionally on linux, it has been tied to the internal
'jiffies' variable, because it had been a cheap and good enough
generator.
For TCP flows on the Internet, 1 ms resolution would be much better
than 4ms or 10ms (HZ=250 or HZ=100 respectively)
For TCP flows in the DC, Google has used usec resolution for more
than two years with great success [1]
Receive size autotuning (DRS) is indeed more precise and converges
faster to optimal window size.
This patch converts tp->tcp_mstamp to a plain u64 value storing
a 1 usec TCP clock.
This choice will allow us to upstream the 1 usec TS option as
discussed in IETF 97.
[1] https://www.ietf.org/proceedings/97/slides/slides-97-tcpm-tcp-options-for-low-latency-00.pdf
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
After this patch, all uses of tcp_time_stamp will require
a change when we introduce 1 ms and/or 1 us TCP TS option.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use tcp_jiffies32 instead of tcp_time_stamp, since
tcp_time_stamp will soon be only used for TCP TS option.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
While working on some recent busy poll changes we found that child sockets
were being instantiated without NAPI ID being set. In our first attempt to
fix it, it was suggested that we should just pull programming the NAPI ID
into the function itself since all callers will need to have it set.
In addition to the NAPI ID change I have dropped the code that was
populating the Rx hash since it was actually being populated in
tcp_get_cookie_sock.
Reported-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Conflicts:
drivers/net/ethernet/broadcom/genet/bcmmii.c
drivers/net/hyperv/netvsc.c
kernel/bpf/hashtab.c
Almost entirely overlapping changes.
Signed-off-by: David S. Miller <davem@davemloft.net>
icsk_ack.lrcvtime has a 0 value at socket creation time.
tcpi_last_data_recv can have bogus value if no payload is ever received.
This patch initializes icsk_ack.lrcvtime for active sessions
in tcp_finish_connect(), and for passive sessions in
tcp_create_openreq_child()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 8a5bd45f6616 (tcp: randomize tcp timestamp offsets for each connection)
randomizes TCP timestamps per connection. After this commit,
there is no guarantee that the timestamps received from the
same destination are monotonically increasing. As a result,
the per-destination timestamp cache in TCP metrics (i.e., tcpm_ts
in struct tcp_metrics_block) is broken and cannot be relied upon.
Remove the per-destination timestamp cache and all related code
paths.
Note that this cache was already broken for caching timestamps of
multiple machines behind a NAT sharing the same address.
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Cc: Lutz Vieweg <lvml@5t9.de>
Cc: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can get SYN with zero tsecr, don't apply offset in this case.
Fixes: ee684b6f28 ("tcp: send packets with a socket timestamp")
Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Small cleanup factorizing code doing the TCP_MAXSEG clamping.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch removes the support of RFC5827 early retransmit (i.e.,
fast recovery on small inflight with <3 dupacks) because it is
subsumed by the new RACK loss detection. More specifically when
RACK receives DUPACKs, it'll arm a reordering timer to start fast
recovery after a quarter of (min)RTT, hence it covers the early
retransmit except RACK does not limit itself to specific inflight
or dupack numbers.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Different namespace application might require fast recycling
TIME-WAIT sockets independently of the host.
Signed-off-by: Haishuang Yan <yanhaishuang@cmss.chinamobile.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
jiffies based timestamps allow for easy inference of number of devices
behind NAT translators and also makes tracking of hosts simpler.
commit ceaa1fef65 ("tcp: adding a per-socket timestamp offset")
added the main infrastructure that is needed for per-connection ts
randomization, in particular writing/reading the on-wire tcp header
format takes the offset into account so rest of stack can use normal
tcp_time_stamp (jiffies).
So only two items are left:
- add a tsoffset for request sockets
- extend the tcp isn generator to also return another 32bit number
in addition to the ISN.
Re-use of ISN generator also means timestamps are still monotonically
increasing for same connection quadruple, i.e. PAWS will still work.
Includes fixes from Eric Dumazet.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit adds code to track whether the delivery rate represented
by each rate_sample was limited by the application.
Upon each transmit, we store in the is_app_limited field in the skb a
boolean bit indicating whether there is a known "bubble in the pipe":
a point in the rate sample interval where the sender was
application-limited, and did not transmit even though the cwnd and
pacing rate allowed it.
This logic marks the flow app-limited on a write if *all* of the
following are true:
1) There is less than 1 MSS of unsent data in the write queue
available to transmit.
2) There is no packet in the sender's queues (e.g. in fq or the NIC
tx queue).
3) The connection is not limited by cwnd.
4) There are no lost packets to retransmit.
The tcp_rate_check_app_limited() code in tcp_rate.c determines whether
the connection is application-limited at the moment. If the flow is
application-limited, it sets the tp->app_limited field. If the flow is
application-limited then that means there is effectively a "bubble" of
silence in the pipe now, and this silence will be reflected in a lower
bandwidth sample for any rate samples from now until we get an ACK
indicating this bubble has exited the pipe: specifically, until we get
an ACK for the next packet we transmit.
When we send every skb we record in scb->tx.is_app_limited whether the
resulting rate sample will be application-limited.
The code in tcp_rate_gen() checks to see when it is safe to mark all
known application-limited bubbles of silence as having exited the
pipe. It does this by checking to see when the delivered count moves
past the tp->app_limited marker. At this point it zeroes the
tp->app_limited marker, as all known bubbles are out of the pipe.
We make room for the tx.is_app_limited bit in the skb by borrowing a
bit from the in_flight field used by NV to record the number of bytes
in flight. The receive window in the TCP header is 16 bits, and the
max receive window scaling shift factor is 14 (RFC 1323). So the max
receive window offered by the TCP protocol is 2^(16+14) = 2^30. So we
only need 30 bits for the tx.in_flight used by NV.
Signed-off-by: Van Jacobson <vanj@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Refactor the TCP min_rtt code to reuse the new win_minmax library in
lib/win_minmax.c to simplify the TCP code.
This is a pure refactor: the functionality is exactly the same. We
just moved the windowed min code to make TCP easier to read and
maintain, and to allow other parts of the kernel to use the windowed
min/max filter code.
Signed-off-by: Van Jacobson <vanj@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Over the years, TCP BDP has increased by several orders of magnitude,
and some people are considering to reach the 2 Gbytes limit.
Even with current window scale limit of 14, ~1 Gbytes maps to ~740,000
MSS.
In presence of packet losses (or reorders), TCP stores incoming packets
into an out of order queue, and number of skbs sitting there waiting for
the missing packets to be received can be in the 10^5 range.
Most packets are appended to the tail of this queue, and when
packets can finally be transferred to receive queue, we scan the queue
from its head.
However, in presence of heavy losses, we might have to find an arbitrary
point in this queue, involving a linear scan for every incoming packet,
throwing away cpu caches.
This patch converts it to a RB tree, to get bounded latencies.
Yaogong wrote a preliminary patch about 2 years ago.
Eric did the rebase, added ofo_last_skb cache, polishing and tests.
Tested with network dropping between 1 and 10 % packets, with good
success (about 30 % increase of throughput in stress tests)
Next step would be to also use an RB tree for the write queue at sender
side ;)
Signed-off-by: Yaogong Wang <wygivan@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Acked-By: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
We want to to make TCP stack preemptible, as draining prequeue
and backlog queues can take lot of time.
Many SNMP updates were assuming that BH (and preemption) was disabled.
Need to convert some __NET_INC_STATS() calls to NET_INC_STATS()
and some __TCP_INC_STATS() to TCP_INC_STATS()
Before using this_cpu_ptr(net->ipv4.tcp_sk) in tcp_v4_send_reset()
and tcp_v4_send_ack(), we add an explicit preempt disabled section.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rename NET_INC_STATS_BH() to __NET_INC_STATS()
and NET_ADD_STATS_BH() to __NET_ADD_STATS()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Attackers like to use SYNFLOOD targeting one 5-tuple, as they
hit a single RX queue (and cpu) on the victim.
If they use random sequence numbers in their SYN, we detect
they do not match the expected window and send back an ACK.
This patch adds a rate limitation, so that the effect of such
attacks is limited to ingress only.
We roughly double our ability to absorb such attacks.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Per RFC4898, they count segments sent/received
containing a positive length data segment (that includes
retransmission segments carrying data). Unlike
tcpi_segs_out/in, tcpi_data_segs_out/in excludes segments
carrying no data (e.g. pure ack).
The patch also updates the segs_in in tcp_fastopen_add_skb()
so that segs_in >= data_segs_in property is kept.
Together with retransmission data, tcpi_data_segs_out
gives a better signal on the rxmit rate.
v6: Rebase on the latest net-next
v5: Eric pointed out that checking skb->len is still needed in
tcp_fastopen_add_skb() because skb can carry a FIN without data.
Hence, instead of open coding segs_in and data_segs_in, tcp_segs_in()
helper is used. Comment is added to the fastopen case to explain why
segs_in has to be reset and tcp_segs_in() has to be called before
__skb_pull().
v4: Add comment to the changes in tcp_fastopen_add_skb()
and also add remark on this case in the commit message.
v3: Add const modifier to the skb parameter in tcp_segs_in()
v2: Rework based on recent fix by Eric:
commit a9d99ce28e ("tcp: fix tcpi_segs_in after connection establishment")
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Cc: Chris Rapier <rapier@psc.edu>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Marcelo Ricardo Leitner <mleitner@redhat.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Several cases of overlapping changes, as well as one instance
(vxlan) of a bug fix in 'net' overlapping with code movement
in 'net-next'.
Signed-off-by: David S. Miller <davem@davemloft.net>
If final packet (ACK) of 3WHS is lost, it appears we do not properly
account the following incoming segment into tcpi_segs_in
While we are at it, starts segs_in with one, to count the SYN packet.
We do not yet count number of SYN we received for a request sock, we
might add this someday.
packetdrill script showing proper behavior after fix :
// Tests tcpi_segs_in when 3rd packet (ACK) of 3WHS is lost
0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+0 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+0 bind(3, ..., ...) = 0
+0 listen(3, 1) = 0
+0 < S 0:0(0) win 32792 <mss 1000,sackOK,nop,nop>
+0 > S. 0:0(0) ack 1 <mss 1460,nop,nop,sackOK>
+.020 < P. 1:1001(1000) ack 1 win 32792
+0 accept(3, ..., ...) = 4
+.000 %{ assert tcpi_segs_in == 2, 'tcpi_segs_in=%d' % tcpi_segs_in }%
Fixes: 2efd055c53 ("tcp: add tcpi_segs_in and tcpi_segs_out to tcp_info")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Hannes points out that when we generate tcp reset for timewait sockets we
pretend we found no socket and pass NULL sk to tcp_vX_send_reset().
Make it cope with inet tw sockets and then provide tw sk.
This makes RSTs appear on correct interface when SO_BINDTODEVICE is used.
Packetdrill test case:
// want default route to be used, we rely on BINDTODEVICE
`ip route del 192.0.2.0/24 via 192.168.0.2 dev tun0`
0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
// test case still works due to BINDTODEVICE
0.001 setsockopt(3, SOL_SOCKET, SO_BINDTODEVICE, "tun0", 4) = 0
0.100...0.200 connect(3, ..., ...) = 0
0.100 > S 0:0(0) <mss 1460,sackOK,nop,nop>
0.200 < S. 0:0(0) ack 1 win 32792 <mss 1460,sackOK,nop,nop>
0.200 > . 1:1(0) ack 1
0.210 close(3) = 0
0.210 > F. 1:1(0) ack 1 win 29200
0.300 < . 1:1(0) ack 2 win 46
// more data while in FIN_WAIT2, expect RST
1.300 < P. 1:1001(1000) ack 1 win 46
// fails without this change -- default route is used
1.301 > R 1:1(0) win 0
Reported-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
For the reasons explained in commit ce1050089c ("tcp/dccp: fix
ireq->pktopts race"), we need to make sure we do not access
req->saved_syn unless we own the request sock.
This fixes races for listeners using TCP_SAVE_SYN option.
Fixes: e994b2f0fb ("tcp: do not lock listener to process SYN packets")
Fixes: 079096f103 ("tcp/dccp: install syn_recv requests into ehash table")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Ying Cai <ycai@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Multiple cpus can process duplicates of incoming ACK messages
matching a SYN_RECV request socket. This is a rare event under
normal operations, but definitely can happen.
Only one must win the race, otherwise corruption would occur.
To fix this without adding new atomic ops, we use logic in
inet_ehash_nolisten() to detect the request was present in the same
ehash bucket where we try to insert the new child.
If request socket was not found, we have to undo the child creation.
This actually removes a spin_lock()/spin_unlock() pair in
reqsk_queue_unlink() for the fast path.
Fixes: e994b2f0fb ("tcp: do not lock listener to process SYN packets")
Fixes: 079096f103 ("tcp/dccp: install syn_recv requests into ehash table")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch is the first half of the RACK loss recovery.
RACK loss recovery uses the notion of time instead
of packet sequence (FACK) or counts (dupthresh). It's inspired by the
previous FACK heuristic in tcp_mark_lost_retrans(): when a limited
transmit (new data packet) is sacked, then current retransmitted
sequence below the newly sacked sequence must been lost,
since at least one round trip time has elapsed.
But it has several limitations:
1) can't detect tail drops since it depends on limited transmit
2) is disabled upon reordering (assumes no reordering)
3) only enabled in fast recovery ut not timeout recovery
RACK (Recently ACK) addresses these limitations with the notion
of time instead: a packet P1 is lost if a later packet P2 is s/acked,
as at least one round trip has passed.
Since RACK cares about the time sequence instead of the data sequence
of packets, it can detect tail drops when later retransmission is
s/acked while FACK or dupthresh can't. For reordering RACK uses a
dynamically adjusted reordering window ("reo_wnd") to reduce false
positives on ever (small) degree of reordering.
This patch implements tcp_advanced_rack() which tracks the
most recent transmission time among the packets that have been
delivered (ACKed or SACKed) in tp->rack.mstamp. This timestamp
is the key to determine which packet has been lost.
Consider an example that the sender sends six packets:
T1: P1 (lost)
T2: P2
T3: P3
T4: P4
T100: sack of P2. rack.mstamp = T2
T101: retransmit P1
T102: sack of P2,P3,P4. rack.mstamp = T4
T205: ACK of P4 since the hole is repaired. rack.mstamp = T101
We need to be careful about spurious retransmission because it may
falsely advance tp->rack.mstamp by an RTT or an RTO, causing RACK
to falsely mark all packets lost, just like a spurious timeout.
We identify spurious retransmission by the ACK's TS echo value.
If TS option is not applicable but the retransmission is acknowledged
less than min-RTT ago, it is likely to be spurious. We refrain from
using the transmission time of these spurious retransmissions.
The second half is implemented in the next patch that marks packet
lost using RACK timestamp.
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>
Kathleen Nichols' algorithm for tracking the minimum RTT of a
data stream over some measurement window. It uses constant space
and constant time per update. Yet it almost always delivers
the same minimum as an implementation that has to keep all
the data in the window. The measurement window is tunable via
sysctl.net.ipv4.tcp_min_rtt_wlen with a default value of 5 minutes.
The algorithm keeps track of the best, 2nd best & 3rd best min
values, maintaining an invariant that the measurement time of
the n'th best >= n-1'th best. It also makes sure that the three
values are widely separated in the time window since that bounds
the worse case error when that data is monotonically increasing
over the window.
Upon getting a new min, we can forget everything earlier because
it has no value - the new min is less than everything else in the
window by definition and it's the most recent. So we restart fresh
on every new min and overwrites the 2nd & 3rd choices. The same
property holds for the 2nd & 3rd best.
Therefore we have to maintain two invariants to maximize the
information in the samples, one on values (1st.v <= 2nd.v <=
3rd.v) and the other on times (now-win <=1st.t <= 2nd.t <= 3rd.t <=
now). These invariants determine the structure of the code
The RTT input to the windowed filter is the minimum RTT measured
from ACK or SACK, or as the last resort from TCP timestamps.
The accessor tcp_min_rtt() returns the minimum RTT seen in the
window. ~0U indicates it is not available. The minimum is 1usec
even if the true RTT is below that.
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>
One 32bit hole is following skc_refcnt, use it.
skc_incoming_cpu can also be an union for request_sock rcv_wnd.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Before recent TCP listener patches, we were updating listener
sk->sk_rxhash before the cloning of master socket.
children sk_rxhash was therefore correct after the normal 3WHS.
But with lockless listener, we no longer dirty/change listener sk_rxhash
as it would be racy.
We need to correctly update the child sk_rxhash, otherwise first data
packet wont hit correct cpu if RFS is used.
Fixes: 079096f103 ("tcp/dccp: install syn_recv requests into ehash table")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Willem de Bruijn <willemb@google.com>
Cc: Tom Herbert <tom@herbertland.com>
Acked-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
