UDP sendmsg() is lockless, so ip_select_ident_segs()
can very well be run from multiple cpus [1]
Convert inet->inet_id to an atomic_t, but implement
a dedicated path for TCP, avoiding cost of a locked
instruction (atomic_add_return())
Note that this patch will cause a trivial merge conflict
because we added inet->flags in net-next tree.
v2: added missing change in
drivers/net/ethernet/chelsio/inline_crypto/chtls/chtls_cm.c
(David Ahern)
[1]
BUG: KCSAN: data-race in __ip_make_skb / __ip_make_skb
read-write to 0xffff888145af952a of 2 bytes by task 7803 on cpu 1:
ip_select_ident_segs include/net/ip.h:542 [inline]
ip_select_ident include/net/ip.h:556 [inline]
__ip_make_skb+0x844/0xc70 net/ipv4/ip_output.c:1446
ip_make_skb+0x233/0x2c0 net/ipv4/ip_output.c:1560
udp_sendmsg+0x1199/0x1250 net/ipv4/udp.c:1260
inet_sendmsg+0x63/0x80 net/ipv4/af_inet.c:830
sock_sendmsg_nosec net/socket.c:725 [inline]
sock_sendmsg net/socket.c:748 [inline]
____sys_sendmsg+0x37c/0x4d0 net/socket.c:2494
___sys_sendmsg net/socket.c:2548 [inline]
__sys_sendmmsg+0x269/0x500 net/socket.c:2634
__do_sys_sendmmsg net/socket.c:2663 [inline]
__se_sys_sendmmsg net/socket.c:2660 [inline]
__x64_sys_sendmmsg+0x57/0x60 net/socket.c:2660
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x41/0xc0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
read to 0xffff888145af952a of 2 bytes by task 7804 on cpu 0:
ip_select_ident_segs include/net/ip.h:541 [inline]
ip_select_ident include/net/ip.h:556 [inline]
__ip_make_skb+0x817/0xc70 net/ipv4/ip_output.c:1446
ip_make_skb+0x233/0x2c0 net/ipv4/ip_output.c:1560
udp_sendmsg+0x1199/0x1250 net/ipv4/udp.c:1260
inet_sendmsg+0x63/0x80 net/ipv4/af_inet.c:830
sock_sendmsg_nosec net/socket.c:725 [inline]
sock_sendmsg net/socket.c:748 [inline]
____sys_sendmsg+0x37c/0x4d0 net/socket.c:2494
___sys_sendmsg net/socket.c:2548 [inline]
__sys_sendmmsg+0x269/0x500 net/socket.c:2634
__do_sys_sendmmsg net/socket.c:2663 [inline]
__se_sys_sendmmsg net/socket.c:2660 [inline]
__x64_sys_sendmmsg+0x57/0x60 net/socket.c:2660
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x41/0xc0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
value changed: 0x184d -> 0x184e
Reported by Kernel Concurrency Sanitizer on:
CPU: 0 PID: 7804 Comm: syz-executor.1 Not tainted 6.5.0-rc6-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/26/2023
==================================================================
Fixes: 23f57406b8 ("ipv4: avoid using shared IP generator for connected sockets")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
IP_RECVERR socket option can now be set/get without locking the socket.
This patch potentially avoid data-races around inet->recverr.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk_getsockopt() runs locklessly. This means sk->sk_priority
can be read while other threads are changing its value.
Other reads also happen without socket lock being held.
Add missing annotations where needed.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk->sk_mark is often read while another thread could change the value.
Fixes: 4a19ec5800 ("[NET]: Introducing socket mark socket option.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Kubernetes[1] is going to stick with /proc/net/tcp for a while.
This commit reduces the scheduling latency introduced by
established_get_first(), similar to commit acffb584cd ("net: diag:
add a scheduling point in inet_diag_dump_icsk()").
In our environment, the scheduling latency affects the performance of
latency-sensitive services like Redis.
Changes in V2 :
- call cond_resched() before checking if a bucket is empty as
suggested by Eric Dumazet
- removed the delay of synchronize_net() from the commit message
[1] https://github.com/google/cadvisor/blob/v0.47.2/container/libcontainer/handler.go#L130
Signed-off-by: Jian Wen <wenjian1@xiaomi.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230711032405.3253025-1-wenjian1@xiaomi.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Under certain circumstances, the tcp receive buffer memory limit
set by autotuning (sk_rcvbuf) is increased due to incoming data
packets as a result of the window not closing when it should be.
This can result in the receive buffer growing all the way up to
tcp_rmem[2], even for tcp sessions with a low BDP.
To reproduce: Connect a TCP session with the receiver doing
nothing and the sender sending small packets (an infinite loop
of socket send() with 4 bytes of payload with a sleep of 1 ms
in between each send()). This will cause the tcp receive buffer
to grow all the way up to tcp_rmem[2].
As a result, a host can have individual tcp sessions with receive
buffers of size tcp_rmem[2], and the host itself can reach tcp_mem
limits, causing the host to go into tcp memory pressure mode.
The fundamental issue is the relationship between the granularity
of the window scaling factor and the number of byte ACKed back
to the sender. This problem has previously been identified in
RFC 7323, appendix F [1].
The Linux kernel currently adheres to never shrinking the window.
In addition to the overallocation of memory mentioned above, the
current behavior is functionally incorrect, because once tcp_rmem[2]
is reached when no remediations remain (i.e. tcp collapse fails to
free up any more memory and there are no packets to prune from the
out-of-order queue), the receiver will drop in-window packets
resulting in retransmissions and an eventual timeout of the tcp
session. A receive buffer full condition should instead result
in a zero window and an indefinite wait.
In practice, this problem is largely hidden for most flows. It
is not applicable to mice flows. Elephant flows can send data
fast enough to "overrun" the sk_rcvbuf limit (in a single ACK),
triggering a zero window.
But this problem does show up for other types of flows. Examples
are websockets and other type of flows that send small amounts of
data spaced apart slightly in time. In these cases, we directly
encounter the problem described in [1].
RFC 7323, section 2.4 [2], says there are instances when a retracted
window can be offered, and that TCP implementations MUST ensure
that they handle a shrinking window, as specified in RFC 1122,
section 4.2.2.16 [3]. All prior RFCs on the topic of tcp window
management have made clear that sender must accept a shrunk window
from the receiver, including RFC 793 [4] and RFC 1323 [5].
This patch implements the functionality to shrink the tcp window
when necessary to keep the right edge within the memory limit by
autotuning (sk_rcvbuf). This new functionality is enabled with
the new sysctl: net.ipv4.tcp_shrink_window
Additional information can be found at:
https://blog.cloudflare.com/unbounded-memory-usage-by-tcp-for-receive-buffers-and-how-we-fixed-it/
[1] https://www.rfc-editor.org/rfc/rfc7323#appendix-F
[2] https://www.rfc-editor.org/rfc/rfc7323#section-2.4
[3] https://www.rfc-editor.org/rfc/rfc1122#page-91
[4] https://www.rfc-editor.org/rfc/rfc793
[5] https://www.rfc-editor.org/rfc/rfc1323
Signed-off-by: Mike Freemon <mfreemon@cloudflare.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow splice to undo the effects of MSG_MORE after prematurely ending a
splice/sendfile due to getting an EOF condition (->splice_read() returned
0) after splice had called sendmsg() with MSG_MORE set when the user didn't
set MSG_MORE.
For UDP, a pending packet will not be emitted if the socket is closed
before it is flushed; with this change, it be flushed by ->splice_eof().
For TCP, it's not clear that MSG_MORE is actually effective.
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/CAHk-=wh=V579PDYvkpnTobCLGczbgxpMgGmmhqiTyE34Cpi5Gg@mail.gmail.com/
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Kuniyuki Iwashima <kuniyu@amazon.com>
cc: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
cc: David Ahern <dsahern@kernel.org>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQTFp0I1jqZrAX+hPRXbK58LschIgwUCZHEm+wAKCRDbK58LschI
gyIKAQCqO7B4sIu8hYVxBTwfHV2tIuXSMSCV4P9e78NUOPcO2QEAvLP/WVSjB0Bm
vpyTKKM22SpZvPe/jSp52j6t20N+qAc=
=HFxD
-----END PGP SIGNATURE-----
Merge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next
Daniel Borkmann says:
====================
pull-request: bpf-next 2023-05-26
We've added 54 non-merge commits during the last 10 day(s) which contain
a total of 76 files changed, 2729 insertions(+), 1003 deletions(-).
The main changes are:
1) Add the capability to destroy sockets in BPF through a new kfunc,
from Aditi Ghag.
2) Support O_PATH fds in BPF_OBJ_PIN and BPF_OBJ_GET commands,
from Andrii Nakryiko.
3) Add capability for libbpf to resize datasec maps when backed via mmap,
from JP Kobryn.
4) Move all the test kfuncs for CI out of the kernel and into bpf_testmod,
from Jiri Olsa.
5) Big batch of xsk selftest improvements to prep for multi-buffer testing,
from Magnus Karlsson.
6) Show the target_{obj,btf}_id in tracing link's fdinfo and dump it
via bpftool, from Yafang Shao.
7) Various misc BPF selftest improvements to work with upcoming LLVM 17,
from Yonghong Song.
8) Extend bpftool to specify netdevice for resolving XDP hints,
from Larysa Zaremba.
9) Document masking in shift operations for the insn set document,
from Dave Thaler.
10) Extend BPF selftests to check xdp_feature support for bond driver,
from Lorenzo Bianconi.
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (54 commits)
bpf: Fix bad unlock balance on freeze_mutex
libbpf: Ensure FD >= 3 during bpf_map__reuse_fd()
libbpf: Ensure libbpf always opens files with O_CLOEXEC
selftests/bpf: Check whether to run selftest
libbpf: Change var type in datasec resize func
bpf: drop unnecessary bpf_capable() check in BPF_MAP_FREEZE command
libbpf: Selftests for resizing datasec maps
libbpf: Add capability for resizing datasec maps
selftests/bpf: Add path_fd-based BPF_OBJ_PIN and BPF_OBJ_GET tests
libbpf: Add opts-based bpf_obj_pin() API and add support for path_fd
bpf: Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
libbpf: Start v1.3 development cycle
bpf: Validate BPF object in BPF_OBJ_PIN before calling LSM
bpftool: Specify XDP Hints ifname when loading program
selftests/bpf: Add xdp_feature selftest for bond device
selftests/bpf: Test bpf_sock_destroy
selftests/bpf: Add helper to get port using getsockname
bpf: Add bpf_sock_destroy kfunc
bpf: Add kfunc filter function to 'struct btf_kfunc_id_set'
bpf: udp: Implement batching for sockets iterator
...
====================
Link: https://lore.kernel.org/r/20230526222747.17775-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When using IPv4/TCP, skb->hash comes from sk->sk_txhash except in
TIME_WAIT and SYN_RECV where it's not set in the reply skb from
ip_send_unicast_reply. Those packets will have a mismatched hash with
others from the same flow as their hashes will be 0. IPv6 does not have
the same issue as the hash is set from the socket txhash in those cases.
This commits sets the hash in the reply skb from ip_send_unicast_reply,
which makes the IPv4 code behaving like IPv6.
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The socket destroy kfunc is used to forcefully terminate sockets from
certain BPF contexts. We plan to use the capability in Cilium
load-balancing to terminate client sockets that continue to connect to
deleted backends. The other use case is on-the-fly policy enforcement
where existing socket connections prevented by policies need to be
forcefully terminated. The kfunc also allows terminating sockets that may
or may not be actively sending traffic.
The kfunc can currently be called only from BPF TCP and UDP iterators
where users can filter, and terminate selected sockets. More
specifically, it can only be called from BPF contexts that ensure
socket locking in order to allow synchronous execution of protocol
specific `diag_destroy` handlers. The previous commit that batches UDP
sockets during iteration facilitated a synchronous invocation of the UDP
destroy callback from BPF context by skipping socket locks in
`udp_abort`. TCP iterator already supported batching of sockets being
iterated. To that end, `tracing_iter_filter` callback filter is added so
that verifier can restrict the kfunc to programs with `BPF_TRACE_ITER`
attach type, and reject other programs.
The kfunc takes `sock_common` type argument, even though it expects, and
casts them to a `sock` pointer. This enables the verifier to allow the
sock_destroy kfunc to be called for TCP with `sock_common` and UDP with
`sock` structs. Furthermore, as `sock_common` only has a subset of
certain fields of `sock`, casting pointer to the latter type might not
always be safe for certain sockets like request sockets, but these have a
special handling in the diag_destroy handlers.
Additionally, the kfunc is defined with `KF_TRUSTED_ARGS` flag to avoid the
cases where a `PTR_TO_BTF_ID` sk is obtained by following another pointer.
eg. getting a sk pointer (may be even NULL) by following another sk
pointer. The pointer socket argument passed in TCP and UDP iterators is
tagged as `PTR_TRUSTED` in {tcp,udp}_reg_info. The TRUSTED arg changes
are contributed by Martin KaFai Lau <martin.lau@kernel.org>.
Signed-off-by: Aditi Ghag <aditi.ghag@isovalent.com>
Link: https://lore.kernel.org/r/20230519225157.760788-8-aditi.ghag@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
This is a preparatory commit to replace `lock_sock_fast` with
`lock_sock`,and facilitate BPF programs executed from the TCP sockets
iterator to be able to destroy TCP sockets using the bpf_sock_destroy
kfunc (implemented in follow-up commits).
Previously, BPF TCP iterator was acquiring the sock lock with BH
disabled. This led to scenarios where the sockets hash table bucket lock
can be acquired with BH enabled in some path versus disabled in other.
In such situation, kernel issued a warning since it thinks that in the
BH enabled path the same bucket lock *might* be acquired again in the
softirq context (BH disabled), which will lead to a potential dead lock.
Since bpf_sock_destroy also happens in a process context, the potential
deadlock warning is likely a false alarm.
Here is a snippet of annotated stack trace that motivated this change:
```
Possible interrupt unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&h->lhash2[i].lock);
local_bh_disable();
lock(&h->lhash2[i].lock);
kernel imagined possible scenario:
local_bh_disable(); /* Possible softirq */
lock(&h->lhash2[i].lock);
*** Potential Deadlock ***
process context:
lock_acquire+0xcd/0x330
_raw_spin_lock+0x33/0x40
------> Acquire (bucket) lhash2.lock with BH enabled
__inet_hash+0x4b/0x210
inet_csk_listen_start+0xe6/0x100
inet_listen+0x95/0x1d0
__sys_listen+0x69/0xb0
__x64_sys_listen+0x14/0x20
do_syscall_64+0x3c/0x90
entry_SYSCALL_64_after_hwframe+0x72/0xdc
bpf_sock_destroy run from iterator:
lock_acquire+0xcd/0x330
_raw_spin_lock+0x33/0x40
------> Acquire (bucket) lhash2.lock with BH disabled
inet_unhash+0x9a/0x110
tcp_set_state+0x6a/0x210
tcp_abort+0x10d/0x200
bpf_prog_6793c5ca50c43c0d_iter_tcp6_server+0xa4/0xa9
bpf_iter_run_prog+0x1ff/0x340
------> lock_sock_fast that acquires sock lock with BH disabled
bpf_iter_tcp_seq_show+0xca/0x190
bpf_seq_read+0x177/0x450
```
Also, Yonghong reported a deadlock for non-listening TCP sockets that
this change resolves. Previously, `lock_sock_fast` held the sock spin
lock with BH which was again being acquired in `tcp_abort`:
```
watchdog: BUG: soft lockup - CPU#0 stuck for 86s! [test_progs:2331]
RIP: 0010:queued_spin_lock_slowpath+0xd8/0x500
Call Trace:
<TASK>
_raw_spin_lock+0x84/0x90
tcp_abort+0x13c/0x1f0
bpf_prog_88539c5453a9dd47_iter_tcp6_client+0x82/0x89
bpf_iter_run_prog+0x1aa/0x2c0
? preempt_count_sub+0x1c/0xd0
? from_kuid_munged+0x1c8/0x210
bpf_iter_tcp_seq_show+0x14e/0x1b0
bpf_seq_read+0x36c/0x6a0
bpf_iter_tcp_seq_show
lock_sock_fast
__lock_sock_fast
spin_lock_bh(&sk->sk_lock.slock);
/* * Fast path return with bottom halves disabled and * sock::sk_lock.slock held.* */
...
tcp_abort
local_bh_disable();
spin_lock(&((sk)->sk_lock.slock)); // from bh_lock_sock(sk)
```
With the switch to `lock_sock`, it calls `spin_unlock_bh` before returning:
```
lock_sock
lock_sock_nested
spin_lock_bh(&sk->sk_lock.slock);
:
spin_unlock_bh(&sk->sk_lock.slock);
```
Acked-by: Yonghong Song <yhs@meta.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Aditi Ghag <aditi.ghag@isovalent.com>
Link: https://lore.kernel.org/r/20230519225157.760788-2-aditi.ghag@isovalent.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Conflicts:
drivers/net/ethernet/freescale/fec_main.c
6ead9c98ca ("net: fec: remove the xdp_return_frame when lack of tx BDs")
144470c88c ("net: fec: using the standard return codes when xdp xmit errors")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When tcp_v4_send_reset() is called with @sk == NULL,
we do not change ctl_sk->sk_priority, which could have been
set from a prior invocation.
Change tcp_v4_send_reset() to set sk_priority and sk_mark
fields before calling ip_send_unicast_reply().
This means tcp_v4_send_reset() and tcp_v4_send_ack()
no longer have to clear ctl_sk->sk_mark after
their call to ip_send_unicast_reply().
Fixes: f6c0f5d209 ("tcp: honor SO_PRIORITY in TIME_WAIT state")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Antoine Tenart <atenart@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently the SYN RTO schedule follows an exponential backoff
scheme, which can be unnecessarily conservative in cases where
there are link failures. In such cases, it's better to
aggressively try to retransmit packets, so it takes routers
less time to find a repath with a working link.
We chose a default value for this sysctl of 4, to follow
the macOS and IOS backoff scheme of 1,1,1,1,1,2,4,8, ...
MacOS and IOS have used this backoff schedule for over
a decade, since before this 2009 IETF presentation
discussed the behavior:
https://www.ietf.org/proceedings/75/slides/tcpm-1.pdf
This commit makes the SYN RTO schedule start with a number of
linear backoffs given by the following sysctl:
* tcp_syn_linear_timeouts
This changes the SYN RTO scheme to be: init_rto_val for
tcp_syn_linear_timeouts, exp backoff starting at init_rto_val
For example if init_rto_val = 1 and tcp_syn_linear_timeouts = 2, our
backoff scheme would be: 1, 1, 1, 2, 4, 8, 16, ...
Signed-off-by: David Morley <morleyd@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Tested-by: David Morley <morleyd@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230509180558.2541885-1-morleyd.kernel@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
While reviewing the udp-iter batching patches, noticed the bpf_iter_tcp
calling sock_put() is incorrect. It should call sock_gen_put instead
because bpf_iter_tcp is iterating the ehash table which has the req sk
and tw sk. This patch replaces all sock_put with sock_gen_put in the
bpf_iter_tcp codepath.
Fixes: 04c7820b77 ("bpf: tcp: Bpf iter batching and lock_sock")
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20230328004232.2134233-1-martin.lau@linux.dev
tcp_poll() reads sk->sk_err without socket lock held/owned.
We should used READ_ONCE() here, and update writers
to use WRITE_ONCE().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This field can be read/written without lock synchronization.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
kfree_rcu(1-arg) should be avoided as much as possible,
since this is only possible from sleepable contexts,
and incurr extra rcu barriers.
I wish the 1-arg variant of kfree_rcu() would
get a distinct name, like kfree_rcu_slow()
to avoid it being abused.
Fixes: 459837b522 ("net/tcp: Disable TCP-MD5 static key on tcp_md5sig_info destruction")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Dmitry Safonov <dima@arista.com>
Link: https://lore.kernel.org/r/20221202052847.2623997-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If the kernel was short on (atomic) memory and failed to allocate it -
don't proceed to creation of request socket. Otherwise the socket would
be unsigned and userspace likely doesn't expect that the TCP is not
MD5-signed anymore.
Signed-off-by: Dmitry Safonov <dima@arista.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
To do that, separate two scenarios:
- where it's the first MD5 key on the system, which means that enabling
of the static key may need to sleep;
- copying of an existing key from a listening socket to the request
socket upon receiving a signed TCP segment, where static key was
already enabled (when the key was added to the listening socket).
Now the life-time of the static branch for TCP-MD5 is until:
- last tcp_md5sig_info is destroyed
- last socket in time-wait state with MD5 key is closed.
Which means that after all sockets with TCP-MD5 keys are gone, the
system gets back the performance of disabled md5-key static branch.
While at here, provide static_key_fast_inc() helper that does ref
counter increment in atomic fashion (without grabbing cpus_read_lock()
on CONFIG_JUMP_LABEL=y). This is needed to add a new user for
a static_key when the caller controls the lifetime of another user.
Signed-off-by: Dmitry Safonov <dima@arista.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a helper to allocate tcp_md5sig_info, that will help later to
do/allocate things when info allocated, once per socket.
Signed-off-by: Dmitry Safonov <dima@arista.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When we call connect() for a socket bound to a wildcard address, we update
saddr locklessly. However, it could result in a data race; another thread
iterating over bhash might see a corrupted address.
Let's update saddr under the bhash bucket's lock.
Fixes: 3df80d9320 ("[DCCP]: Introduce DCCPv6")
Fixes: 7c657876b6 ("[DCCP]: Initial implementation")
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When connect() is called on a socket bound to the wildcard address,
we change the socket's saddr to a local address. If the socket
fails to connect() to the destination, we have to reset the saddr.
However, when an error occurs after inet_hash6?_connect() in
(dccp|tcp)_v[46]_conect(), we forget to reset saddr and leave
the socket bound to the address.
From the user's point of view, whether saddr is reset or not varies
with errno. Let's fix this inconsistent behaviour.
Note that after this patch, the repro [0] will trigger the WARN_ON()
in inet_csk_get_port() again, but this patch is not buggy and rather
fixes a bug papering over the bhash2's bug for which we need another
fix.
For the record, the repro causes -EADDRNOTAVAIL in inet_hash6_connect()
by this sequence:
s1 = socket()
s1.setsockopt(SOL_SOCKET, SO_REUSEADDR, 1)
s1.bind(('127.0.0.1', 10000))
s1.sendto(b'hello', MSG_FASTOPEN, (('127.0.0.1', 10000)))
# or s1.connect(('127.0.0.1', 10000))
s2 = socket()
s2.setsockopt(SOL_SOCKET, SO_REUSEADDR, 1)
s2.bind(('0.0.0.0', 10000))
s2.connect(('127.0.0.1', 10000)) # -EADDRNOTAVAIL
s2.listen(32) # WARN_ON(inet_csk(sk)->icsk_bind2_hash != tb2);
[0]: https://syzkaller.appspot.com/bug?extid=015d756bbd1f8b5c8f09
Fixes: 3df80d9320 ("[DCCP]: Introduce DCCPv6")
Fixes: 7c657876b6 ("[DCCP]: Initial implementation")
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The value of 'st->state' has been verified as "TCP_SEQ_STATE_LISTENING",
it's unnecessary to assign TCP_SEQ_STATE_LISTENING to it, so we can remove it.
Signed-off-by: xu xin <xu.xin16@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
Congestion control algorithms track PLB state and cause the connection
to trigger a path change when either of the 2 conditions is satisfied:
- No packets are in flight and (# consecutive congested rounds >=
sysctl_tcp_plb_idle_rehash_rounds)
- (# consecutive congested rounds >= sysctl_tcp_plb_rehash_rounds)
A round (RTT) is marked as congested when congestion signal
(ECN ce_ratio) over an RTT is greater than sysctl_tcp_plb_cong_thresh.
In the event of RTO, PLB (via tcp_write_timeout()) triggers a path
change and disables congestion-triggered path changes for random time
between (sysctl_tcp_plb_suspend_rto_sec, 2*sysctl_tcp_plb_suspend_rto_sec)
to avoid hopping onto the "connectivity blackhole". RTO-triggered
path changes can still happen during this cool-off period.
Signed-off-by: Mubashir Adnan Qureshi <mubashirq@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
PLB (Protective Load Balancing) is a host based mechanism for load
balancing across switch links. It leverages congestion signals(e.g. ECN)
from transport layer to randomly change the path of the connection
experiencing congestion. PLB changes the path of the connection by
changing the outgoing IPv6 flow label for IPv6 connections (implemented
in Linux by calling sk_rethink_txhash()). Because of this implementation
mechanism, PLB can currently only work for IPv6 traffic. For more
information, see the SIGCOMM 2022 paper:
https://doi.org/10.1145/3544216.3544226
This commit adds new sysctl knobs and sets their default values for
TCP PLB.
Signed-off-by: Mubashir Adnan Qureshi <mubashirq@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The type of sk_rcvbuf and sk_sndbuf in struct sock is int, and
in tcp_add_backlog(), the variable limit is caculated by adding
sk_rcvbuf, sk_sndbuf and 64 * 1024, it may exceed the max value
of int and overflow. This patch reduces the limit budget by
halving the sndbuf to solve this issue since ACK packets are much
smaller than the payload.
Fixes: c9c3321257 ("tcp: add tcp_add_backlog()")
Signed-off-by: Lu Wei <luwei32@huawei.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
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.
This means we cannot use the global sk->sk_prot->h.hashinfo
to fetch a TCP hashinfo.
Instead, set NULL to sk->sk_prot->h.hashinfo for TCP and get
a proper hashinfo from net->ipv4.tcp_death_row.hashinfo.
Note that we need not use sk->sk_prot->h.hashinfo if DCCP is
disabled.
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>
Because per host rate limiting has been proven problematic (side channel
attacks can be based on it), per host rate limiting of challenge acks ideally
should be per netns and turned off by default.
This is a long due followup of following commits:
083ae30828 ("tcp: enable per-socket rate limiting of all 'challenge acks'")
f2b2c582e8 ("tcp: mitigate ACK loops for connections as tcp_sock")
75ff39ccc1 ("tcp: make challenge acks less predictable")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jason Baron <jbaron@akamai.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The current bind hashtable (bhash) is hashed by port only.
In the socket bind path, we have to check for bind conflicts by
traversing the specified port's inet_bind_bucket while holding the
hashbucket's spinlock (see inet_csk_get_port() and
inet_csk_bind_conflict()). In instances where there are tons of
sockets hashed to the same port at different addresses, the bind
conflict check is time-intensive and can cause softirq cpu lockups,
as well as stops new tcp connections since __inet_inherit_port()
also contests for the spinlock.
This patch adds a second bind table, bhash2, that hashes by
port and sk->sk_rcv_saddr (ipv4) and sk->sk_v6_rcv_saddr (ipv6).
Searching the bhash2 table leads to significantly faster conflict
resolution and less time holding the hashbucket spinlock.
Please note a few things:
* There can be the case where the a socket's address changes after it
has been bound. There are two cases where this happens:
1) The case where there is a bind() call on INADDR_ANY (ipv4) or
IPV6_ADDR_ANY (ipv6) and then a connect() call. The kernel will
assign the socket an address when it handles the connect()
2) In inet_sk_reselect_saddr(), which is called when rebuilding the
sk header and a few pre-conditions are met (eg rerouting fails).
In these two cases, we need to update the bhash2 table by removing the
entry for the old address, and add a new entry reflecting the updated
address.
* The bhash2 table must have its own lock, even though concurrent
accesses on the same port are protected by the bhash lock. Bhash2 must
have its own lock to protect against cases where sockets on different
ports hash to different bhash hashbuckets but to the same bhash2
hashbucket.
This brings up a few stipulations:
1) When acquiring both the bhash and the bhash2 lock, the bhash2 lock
will always be acquired after the bhash lock and released before the
bhash lock is released.
2) There are no nested bhash2 hashbucket locks. A bhash2 lock is always
acquired+released before another bhash2 lock is acquired+released.
* The bhash table cannot be superseded by the bhash2 table because for
bind requests on INADDR_ANY (ipv4) or IPV6_ADDR_ANY (ipv6), every socket
bound to that port must be checked for a potential conflict. The bhash
table is the only source of port->socket associations.
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
While reading sysctl_tcp_reflect_tos, it can be changed concurrently.
Thus, we need to add READ_ONCE() to its readers.
Fixes: ac8f1710c1 ("tcp: reflect tos value received in SYN to the socket")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Wei Wang <weiwan@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jakub Kicinski