| Age | Commit message (Collapse) | Author |
|---|
|
Add a new SKB drop reason (SKB_DROP_REASON_PFMEMALLOC) to track packets
dropped due to memory pressure. In production environments, we've observed
memory exhaustion reported by memory layer stack traces, but these drops
were not properly tracked in the SKB drop reason infrastructure.
While most network code paths now properly report pfmemalloc drops, some
protocol-specific socket implementations still use sk_filter() without
drop reason tracking:
- Bluetooth L2CAP sockets
- CAIF sockets
- IUCV sockets
- Netlink sockets
- SCTP sockets
- Unix domain sockets
These remaining cases represent less common paths and could be converted
in a follow-up patch if needed. The current implementation provides
significantly improved observability into memory pressure events in the
network stack, especially for key protocols like TCP and UDP, helping to
diagnose problems in production environments.
Reported-by: Matt Fleming <mfleming@cloudflare.com>
Signed-off-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://patch.msgid.link/175268316579.2407873.11634752355644843509.stgit@firesoul
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next
Martin KaFai Lau says:
====================
pull-request: bpf-next 2025-07-17
We've added 13 non-merge commits during the last 20 day(s) which contain
a total of 4 files changed, 712 insertions(+), 84 deletions(-).
The main changes are:
1) Avoid skipping or repeating a sk when using a TCP bpf_iter,
from Jordan Rife.
2) Clarify the driver requirement on using the XDP metadata,
from Song Yoong Siang
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next:
doc: xdp: Clarify driver implementation for XDP Rx metadata
selftests/bpf: Add tests for bucket resume logic in established sockets
selftests/bpf: Create iter_tcp_destroy test program
selftests/bpf: Create established sockets in socket iterator tests
selftests/bpf: Make ehash buckets configurable in socket iterator tests
selftests/bpf: Allow for iteration over multiple states
selftests/bpf: Allow for iteration over multiple ports
selftests/bpf: Add tests for bucket resume logic in listening sockets
bpf: tcp: Avoid socket skips and repeats during iteration
bpf: tcp: Use bpf_tcp_iter_batch_item for bpf_tcp_iter_state batch items
bpf: tcp: Get rid of st_bucket_done
bpf: tcp: Make sure iter->batch always contains a full bucket snapshot
bpf: tcp: Make mem flags configurable through bpf_iter_tcp_realloc_batch
====================
Link: https://patch.msgid.link/20250717191731.4142326-1-martin.lau@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Replace the offset-based approach for tracking progress through a bucket
in the TCP table with one based on socket cookies. Remember the cookies
of unprocessed sockets from the last batch and use this list to
pick up where we left off or, in the case that the next socket
disappears between reads, find the first socket after that point that
still exists in the bucket and resume from there.
This approach guarantees that all sockets that existed when iteration
began and continue to exist throughout will be visited exactly once.
Sockets that are added to the table during iteration may or may not be
seen, but if they are they will be seen exactly once.
Signed-off-by: Jordan Rife <jordan@jrife.io>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
|
|
Prepare for the next patch that tracks cookies between iterations by
converting struct sock **batch to union bpf_tcp_iter_batch_item *batch
inside struct bpf_tcp_iter_state.
Signed-off-by: Jordan Rife <jordan@jrife.io>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
|
|
Get rid of the st_bucket_done field to simplify TCP iterator state and
logic. Before, st_bucket_done could be false if bpf_iter_tcp_batch
returned a partial batch; however, with the last patch ("bpf: tcp: Make
sure iter->batch always contains a full bucket snapshot"),
st_bucket_done == true is equivalent to iter->cur_sk == iter->end_sk.
Signed-off-by: Jordan Rife <jordan@jrife.io>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
|
|
Require that iter->batch always contains a full bucket snapshot. This
invariant is important to avoid skipping or repeating sockets during
iteration when combined with the next few patches. Before, there were
two cases where a call to bpf_iter_tcp_batch may only capture part of a
bucket:
1. When bpf_iter_tcp_realloc_batch() returns -ENOMEM.
2. When more sockets are added to the bucket while calling
bpf_iter_tcp_realloc_batch(), making the updated batch size
insufficient.
In cases where the batch size only covers part of a bucket, it is
possible to forget which sockets were already visited, especially if we
have to process a bucket in more than two batches. This forces us to
choose between repeating or skipping sockets, so don't allow this:
1. Stop iteration and propagate -ENOMEM up to userspace if reallocation
fails instead of continuing with a partial batch.
2. Try bpf_iter_tcp_realloc_batch() with GFP_USER just as before, but if
we still aren't able to capture the full bucket, call
bpf_iter_tcp_realloc_batch() again while holding the bucket lock to
guarantee the bucket does not change. On the second attempt use
GFP_NOWAIT since we hold onto the spin lock.
I did some manual testing to exercise the code paths where GFP_NOWAIT is
used and where ERR_PTR(err) is returned. I used the realloc test cases
included later in this series to trigger a scenario where a realloc
happens inside bpf_iter_tcp_batch and made a small code tweak to force
the first realloc attempt to allocate a too-small batch, thus requiring
another attempt with GFP_NOWAIT. Some printks showed both reallocs with
the tests passing:
Jun 27 00:00:53 crow kernel: again GFP_USER
Jun 27 00:00:53 crow kernel: again GFP_NOWAIT
Jun 27 00:00:53 crow kernel: again GFP_USER
Jun 27 00:00:53 crow kernel: again GFP_NOWAIT
With this setup, I also forced each of the bpf_iter_tcp_realloc_batch
calls to return -ENOMEM to ensure that iteration ends and that the
read() in userspace fails.
Signed-off-by: Jordan Rife <jordan@jrife.io>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
|
|
Prepare for the next patch which needs to be able to choose either
GFP_USER or GFP_NOWAIT for calls to bpf_iter_tcp_realloc_batch.
Signed-off-by: Jordan Rife <jordan@jrife.io>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
|
|
Use the new helpers as a first step to deal with
potential dst->dev races.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20250630121934.3399505-8-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
____cacheline_aligned_in_smp attribute only makes sure to align
a field to a cache line. It does not prevent the linker to use
the remaining of the cache line for other variables, causing
potential false sharing.
Move tcp_memory_allocated into a dedicated cache line.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/20250630093540.3052835-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Now inet_rtx_syn_ack() is only used by TCP, it can directly
call tcp_rtx_synack() instead of using an indirect call
to req->rsk_ops->rtx_syn_ack().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20250626153017.2156274-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Difference between sock_i_uid() and sk_uid() is that
after sock_orphan(), sock_i_uid() returns GLOBAL_ROOT_UID
while sk_uid() returns the last cached sk->sk_uid value.
None of sock_i_uid() callers care about this.
Use sk_uid() which is much faster and inlined.
Note that diag/dump users are calling sock_i_ino() and
can not see the full benefit yet.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Lorenzo Colitti <lorenzo@google.com>
Reviewed-by: Maciej Żenczykowski <maze@google.com>
Link: https://patch.msgid.link/20250620133001.4090592-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Last change happened in 2018 with commit c73e5807e4f6
("tcp: tsq: no longer use limit_output_bytes for paced flows")
Modern NIC speeds got a 4x increase since then.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250513193919.1089692-10-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Devices in the networking path, such as firewalls, NATs, or routers, which
can perform SNAT or DNAT, use addresses from their own limited address
pools to masquerade the source address during forwarding, causing PAWS
verification to fail more easily.
Currently, packet loss statistics for PAWS can only be viewed through MIB,
which is a global metric and cannot be precisely obtained through tracing
to get the specific 4-tuple of the dropped packet. In the past, we had to
use kprobe ret to retrieve relevant skb information from
tcp_timewait_state_process().
We add a drop_reason pointer, similar to what previous commit does:
commit e34100c2ecbb ("tcp: add a drop_reason pointer to tcp_check_req()")
This commit addresses the PAWSESTABREJECTED case and also sets the
corresponding drop reason.
We use 'pwru' to test.
Before this commit:
''''
./pwru 'port 9999'
2025/04/07 13:40:19 Listening for events..
TUPLE FUNC
172.31.75.115:12345->172.31.75.114:9999(tcp) sk_skb_reason_drop(SKB_DROP_REASON_NOT_SPECIFIED)
'''
After this commit:
'''
./pwru 'port 9999'
2025/04/07 13:51:34 Listening for events..
TUPLE FUNC
172.31.75.115:12345->172.31.75.114:9999(tcp) sk_skb_reason_drop(SKB_DROP_REASON_TCP_RFC7323_TW_PAWS)
'''
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250409112614.16153-2-jiayuan.chen@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
icsk->icsk_timeout can be replaced by icsk->icsk_retransmit_timer.expires
This saves 8 bytes in TCP/DCCP sockets and helps for better cache locality.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250324203607.703850-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
inet_connection_sock_af_ops.addr2sockaddr() hasn't been used at all
in the git era.
$ git grep addr2sockaddr $(git rev-list HEAD | tail -n 1)
Let's remove it.
Note that there was a 4 bytes hole after sockaddr_len and now it's
6 bytes, so the binary layout is not changed.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250318060112.3729-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
ECN bits in TOS are always cleared when sending in ACKs in TW. Clearing
them is problematic for TCP flows that used Accurate ECN because ECN bits
decide which service queue the packet is placed into (L4S vs Classic).
Effectively, TW ACKs are always downgraded from L4S to Classic queue
which might impact, e.g., delay the ACK will experience on the path
compared with the other packets of the flow.
Change the TW ACK sending code to differentiate:
- In tcp_v4_send_reset(), commit ba9e04a7ddf4f ("ip: fix tos reflection
in ack and reset packets") cleans ECN bits for TW reset and this is
not affected.
- In tcp_v4_timewait_ack(), ECN bits for all TW ACKs are cleaned. But now
only ECN bits of ACKs for oow data or paws_reject are cleaned, and ECN
bits of other ACKs will not be cleaned.
- In tcp_v4_reqsk_send_ack(), commit 66b13d99d96a1 ("ipv4: tcp: fix TOS
value in ACK messages sent from TIME_WAIT") did not clean ECN bits of
ACKs for oow data or paws_reject. But now the ECN bits rae cleaned for
these ACKs.
Signed-off-by: Ilpo Järvinen <ij@kernel.org>
Signed-off-by: Chia-Yu Chang <chia-yu.chang@nokia-bell-labs.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
AE flag needs to be preserved for AccECN.
Signed-off-by: Ilpo Järvinen <ij@kernel.org>
Signed-off-by: Chia-Yu Chang <chia-yu.chang@nokia-bell-labs.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
With AccECN, there's one additional TCP flag to be used (AE)
and ACE field that overloads the definition of AE, CWR, and
ECE flags. As tcp_flags was previously only 1 byte, the
byte-order stuff needs to be added to it's handling.
Signed-off-by: Ilpo Järvinen <ij@kernel.org>
Signed-off-by: Chia-Yu Chang <chia-yu.chang@nokia-bell-labs.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
After commit 8d52da23b6c6 ("tcp: Defer ts_recent changes
until req is owned"), req->ts_recent is not changed anymore.
It is set once in tcp_openreq_init(), bpf_sk_assign_tcp_reqsk()
or cookie_tcp_reqsk_alloc() before the req can be seen by other
cpus/threads.
This completes the revert of eba20811f326 ("tcp: annotate
data-races around tcp_rsk(req)->ts_recent").
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Wang Hai <wanghai38@huawei.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250301201424.2046477-6-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
TCP uses of dev_net() are under RCU protection, change them
to dev_net_rcu() to get LOCKDEP support.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250301201424.2046477-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
We want to add new drop reasons for packets dropped in 3WHS in the
following patches.
tcp_rcv_state_process() has to set reason to TCP_FASTOPEN,
because tcp_check_req() will conditionally overwrite the drop_reason.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250301201424.2046477-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Cross-merge networking fixes after downstream PR (net-6.14-rc4).
No conflicts or adjacent changes.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Xiumei reported hitting the WARN in xfrm6_tunnel_net_exit while
running tests that boil down to:
- create a pair of netns
- run a basic TCP test over ipcomp6
- delete the pair of netns
The xfrm_state found on spi_byaddr was not deleted at the time we
delete the netns, because we still have a reference on it. This
lingering reference comes from a secpath (which holds a ref on the
xfrm_state), which is still attached to an skb. This skb is not
leaked, it ends up on sk_receive_queue and then gets defer-free'd by
skb_attempt_defer_free.
The problem happens when we defer freeing an skb (push it on one CPU's
defer_list), and don't flush that list before the netns is deleted. In
that case, we still have a reference on the xfrm_state that we don't
expect at this point.
We already drop the skb's dst in the TCP receive path when it's no
longer needed, so let's also drop the secpath. At this point,
tcp_filter has already called into the LSM hooks that may require the
secpath, so it should not be needed anymore. However, in some of those
places, the MPTCP extension has just been attached to the skb, so we
cannot simply drop all extensions.
Fixes: 68822bdf76f1 ("net: generalize skb freeing deferral to per-cpu lists")
Reported-by: Xiumei Mu <xmu@redhat.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/5055ba8f8f72bdcb602faa299faca73c280b7735.1739743613.git.sd@queasysnail.net
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Current inet_sock_set_state trace from inet_csk_clone_lock() is missing
many details :
... sock:inet_sock_set_state: family=AF_INET6 protocol=IPPROTO_TCP \
sport=4901 dport=0 \
saddr=127.0.0.6 daddr=0.0.0.0 \
saddrv6=:: daddrv6=:: \
oldstate=TCP_LISTEN newstate=TCP_SYN_RECV
Only the sport gives the listener port, no other parts of the n-tuple are correct.
In this patch, I initialize relevant fields of the new socket before
calling inet_sk_set_state(newsk, TCP_SYN_RECV).
We now have a trace including all the source/destination bits.
... sock:inet_sock_set_state: family=AF_INET6 protocol=IPPROTO_TCP \
sport=4901 dport=47648 \
saddr=127.0.0.6 daddr=127.0.0.6 \
saddrv6=2002:a05:6830:1f85:: daddrv6=2001:4860:f803:65::3 \
oldstate=TCP_LISTEN newstate=TCP_SYN_RECV
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://patch.msgid.link/20250212131328.1514243-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Use EXPORT_IPV6_MOD[_GPL]() for symbols that don't need
to be exported unless CONFIG_IPV6=m
tcp_hashinfo and tcp_openreq_init_rwin() are no longer
used from any module anyway.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Mateusz Polchlopek <mateusz.polchlopek@intel.com>
Link: https://patch.msgid.link/20250212132418.1524422-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Previous patch added a TCP_RTO_MAX_MS socket option
to tune a TCP socket max RTO value.
Many setups prefer to change a per netns sysctl.
This patch adds /proc/sys/net/ipv4/tcp_rto_max_ms
Its initial value is 120000 (120 seconds).
Keep in mind that a decrease of tcp_rto_max_ms
means shorter overall timeouts, unless tcp_retries2
sysctl is increased.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Currently, TCP stack uses a constant (120 seconds)
to limit the RTO value exponential growth.
Some applications want to set a lower value.
Add TCP_RTO_MAX_MS socket option to set a value (in ms)
between 1 and 120 seconds.
It is discouraged to change the socket rto max on a live
socket, as it might lead to unexpected disconnects.
Following patch is adding a netns sysctl to control the
default value at socket creation time.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
In order to reduce TCP_RTO_MAX occurrences, replace:
inet_csk_reset_xmit_timer(sk, what, when, TCP_RTO_MAX)
With:
tcp_reset_xmit_timer(sk, what, when, false);
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Cross-merge networking fixes after downstream PR (net-6.13-rc7).
Conflicts:
a42d71e322a8 ("net_sched: sch_cake: Add drop reasons")
737d4d91d35b ("sched: sch_cake: add bounds checks to host bulk flow fairness counts")
Adjacent changes:
drivers/net/ethernet/meta/fbnic/fbnic.h
3a856ab34726 ("eth: fbnic: add IRQ reuse support")
95978931d55f ("eth: fbnic: Revert "eth: fbnic: Add hardware monitoring support via HWMON interface"")
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
This is a follow-up to 3c5b4d69c358 ("net: annotate data-races around
sk->sk_mark"). sk->sk_mark can be read and written without holding
the socket lock. IPv6 equivalent is already covered with READ_ONCE()
annotation in tcp_v6_send_response().
Fixes: 3c5b4d69c358 ("net: annotate data-races around sk->sk_mark")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/f459d1fc44f205e13f6d8bdca2c8bfb9902ffac9.1736244569.git.daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Today we have a hardcoded delay of 1 sec before a TIME-WAIT socket can be
reused by reopening a connection. This is a safe choice based on an
assumption that the other TCP timestamp clock frequency, which is unknown
to us, may be as low as 1 Hz (RFC 7323, section 5.4).
However, this means that in the presence of short lived connections with an
RTT of couple of milliseconds, the time during which a 4-tuple is blocked
from reuse can be orders of magnitude longer that the connection lifetime.
Combined with a reduced pool of ephemeral ports, when using
IP_LOCAL_PORT_RANGE to share an egress IP address between hosts [1], the
long TIME-WAIT reuse delay can lead to port exhaustion, where all available
4-tuples are tied up in TIME-WAIT state.
Turn the reuse delay into a per-netns setting so that sysadmins can make
more aggressive assumptions about remote TCP timestamp clock frequency and
shorten the delay in order to allow connections to reincarnate faster.
Note that applications can completely bypass the TIME-WAIT delay protection
already today by locking the local port with bind() before connecting. Such
immediate connection reuse may result in PAWS failing to detect old
duplicate segments, leaving us with just the sequence number check as a
safety net.
This new configurable offers a trade off where the sysadmin can balance
between the risk of PAWS detection failing to act versus exhausting ports
by having sockets tied up in TIME-WAIT state for too long.
[1] https://lpc.events/event/16/contributions/1349/
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Link: https://patch.msgid.link/20241209-jakub-krn-909-poc-msec-tw-tstamp-v2-2-66aca0eed03e@cloudflare.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Prepare ground for TIME-WAIT socket reuse with subsecond delay.
Today the last TS.Recent update timestamp, recorded in seconds and stored
tp->ts_recent_stamp and tw->tw_ts_recent_stamp fields, has two purposes.
Firstly, it is used to track the age of the last recorded TS.Recent value
to detect when that value becomes outdated due to potential wrap-around of
the other TCP timestamp clock (RFC 7323, section 5.5).
For this purpose a second-based timestamp is completely sufficient as even
in the worst case scenario of a peer using a high resolution microsecond
timestamp, the wrap-around interval is ~36 minutes long.
Secondly, it serves as a threshold value for allowing TIME-WAIT socket
reuse. A TIME-WAIT socket can be reused only once the virtual 1 Hz clock,
ktime_get_seconds, is past the TS.Recent update timestamp.
The purpose behind delaying the TIME-WAIT socket reuse is to wait for the
other TCP timestamp clock to tick at least once before reusing the
connection. It is only then that the PAWS mechanism for the reopened
connection can detect old duplicate segments from the previous connection
incarnation (RFC 7323, appendix B.2).
In this case using a timestamp with second resolution not only blocks the
way toward allowing faster TIME-WAIT reuse after shorter subsecond delay,
but also makes it impossible to reliably delay TW reuse by one second.
As Eric Dumazet has pointed out [1], due to timestamp rounding, the TW
reuse delay will actually be between (0, 1] seconds, and 0.5 seconds on
average. We delay TW reuse for one full second only when last TS.Recent
update coincides with our virtual 1 Hz clock tick.
Considering the above, introduce a dedicated field to store a millisecond
timestamp of transition into the TIME-WAIT state. Place it in an existing
4-byte hole inside inet_timewait_sock structure to avoid an additional
memory cost.
Use the new timestamp to (i) reliably delay TIME-WAIT reuse by one second,
and (ii) prepare for configurable subsecond reuse delay in the subsequent
change.
We assume here that a full one second delay was the original intention in
[2] because it accounts for the worst case scenario of the other TCP using
the slowest recommended 1 Hz timestamp clock.
A more involved alternative would be to change the resolution of the last
TS.Recent update timestamp, tw->tw_ts_recent_stamp, to milliseconds.
[1] https://lore.kernel.org/netdev/CANn89iKB4GFd8sVzCbRttqw_96o3i2wDhX-3DraQtsceNGYwug@mail.gmail.com/
[2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b8439924316d5bcb266d165b93d632a4b4b859af
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Link: https://patch.msgid.link/20241209-jakub-krn-909-poc-msec-tw-tstamp-v2-1-66aca0eed03e@cloudflare.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Under CONFIG_PROVE_RCU_LIST + CONFIG_RCU_EXPERT
hlist_for_each_entry_rcu() provides very helpful splats, which help
to find possible issues. I missed CONFIG_RCU_EXPERT=y in my testing
config the same as described in
a3e4bf7f9675 ("configs/debug: make sure PROVE_RCU_LIST=y takes effect").
The fix itself is trivial: add the very same lockdep annotations
as were used to dereference ao_info from the socket.
Reported-by: Jakub Kicinski <kuba@kernel.org>
Closes: https://lore.kernel.org/netdev/20241028152645.35a8be66@kernel.org/
Signed-off-by: Dmitry Safonov <0x7f454c46@gmail.com>
Link: https://patch.msgid.link/20241030-tcp-ao-hlist-lockdep-annotate-v1-1-bf641a64d7c6@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
ip_send_unicast_reply() send orphaned 'control packets'.
These are RST packets and also ACK packets sent from TIME_WAIT.
Some eBPF programs would prefer to have a meaningful skb->sk
pointer as much as possible.
This means that TCP can now attach TIME_WAIT sockets to outgoing
skbs.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Brian Vazquez <brianvv@google.com>
Link: https://patch.msgid.link/20241010174817.1543642-6-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
icsk->icsk_pending can be read locklessly already.
Following patch in the series will add another lockless read.
Add smp_load_acquire() and smp_store_release() annotations
because following patch will add a test in tcp_write_timer(),
and READ_ONCE()/WRITE_ONCE() alone would possibly lead to races.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241002173042.917928-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
In tcp_recvmsg_locked(), detect if the skb being received by the user
is a devmem skb. In this case - if the user provided the MSG_SOCK_DEVMEM
flag - pass it to tcp_recvmsg_devmem() for custom handling.
tcp_recvmsg_devmem() copies any data in the skb header to the linear
buffer, and returns a cmsg to the user indicating the number of bytes
returned in the linear buffer.
tcp_recvmsg_devmem() then loops over the unaccessible devmem skb frags,
and returns to the user a cmsg_devmem indicating the location of the
data in the dmabuf device memory. cmsg_devmem contains this information:
1. the offset into the dmabuf where the payload starts. 'frag_offset'.
2. the size of the frag. 'frag_size'.
3. an opaque token 'frag_token' to return to the kernel when the buffer
is to be released.
The pages awaiting freeing are stored in the newly added
sk->sk_user_frags, and each page passed to userspace is get_page()'d.
This reference is dropped once the userspace indicates that it is
done reading this page. All pages are released when the socket is
destroyed.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Kaiyuan Zhang <kaiyuanz@google.com>
Signed-off-by: Mina Almasry <almasrymina@google.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20240910171458.219195-10-almasrymina@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
No lock protects tcp tw fields.
tcptw->tw_rcv_nxt can be changed from twsk_rcv_nxt_update()
while other threads might read this field.
Add READ_ONCE()/WRITE_ONCE() annotations, and make sure
tcp_timewait_state_process() reads tcptw->tw_rcv_nxt only once.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Link: https://patch.msgid.link/20240827015250.3509197-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Using a volatile qualifier for a specific struct field is unusual.
Use instead READ_ONCE()/WRITE_ONCE() where necessary.
tcp_timewait_state_process() can change tw_substate while other
threads are reading this field.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Link: https://patch.msgid.link/20240827015250.3509197-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
We found that one close-wait socket was reset by the other side
due to a new connection reusing the same port which is beyond our
expectation, so we have to investigate the underlying reason.
The following experiment is conducted in the test environment. We
limit the port range from 40000 to 40010 and delay the time to close()
after receiving a fin from the active close side, which can help us
easily reproduce like what happened in production.
Here are three connections captured by tcpdump:
127.0.0.1.40002 > 127.0.0.1.9999: Flags [S], seq 2965525191
127.0.0.1.9999 > 127.0.0.1.40002: Flags [S.], seq 2769915070
127.0.0.1.40002 > 127.0.0.1.9999: Flags [.], ack 1
127.0.0.1.40002 > 127.0.0.1.9999: Flags [F.], seq 1, ack 1
// a few seconds later, within 60 seconds
127.0.0.1.40002 > 127.0.0.1.9999: Flags [S], seq 2965590730
127.0.0.1.9999 > 127.0.0.1.40002: Flags [.], ack 2
127.0.0.1.40002 > 127.0.0.1.9999: Flags [R], seq 2965525193
// later, very quickly
127.0.0.1.40002 > 127.0.0.1.9999: Flags [S], seq 2965590730
127.0.0.1.9999 > 127.0.0.1.40002: Flags [S.], seq 3120990805
127.0.0.1.40002 > 127.0.0.1.9999: Flags [.], ack 1
As we can see, the first flow is reset because:
1) client starts a new connection, I mean, the second one
2) client tries to find a suitable port which is a timewait socket
(its state is timewait, substate is fin_wait2)
3) client occupies that timewait port to send a SYN
4) server finds a corresponding close-wait socket in ehash table,
then replies with a challenge ack
5) client sends an RST to terminate this old close-wait socket.
I don't think the port selection algo can choose a FIN_WAIT2 socket
when we turn on tcp_tw_reuse because on the server side there
remain unread data. In some cases, if one side haven't call close() yet,
we should not consider it as expendable and treat it at will.
Even though, sometimes, the server isn't able to call close() as soon
as possible like what we expect, it can not be terminated easily,
especially due to a second unrelated connection happening.
After this patch, we can see the expected failure if we start a
connection when all the ports are occupied in fin_wait2 state:
"Ncat: Cannot assign requested address."
Reported-by: Jade Dong <jadedong@tencent.com>
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20240823001152.31004-1-kerneljasonxing@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Its possible that two threads call tcp_sk_exit_batch() concurrently,
once from the cleanup_net workqueue, once from a task that failed to clone
a new netns. In the latter case, error unwinding calls the exit handlers
in reverse order for the 'failed' netns.
tcp_sk_exit_batch() calls tcp_twsk_purge().
Problem is that since commit b099ce2602d8 ("net: Batch inet_twsk_purge"),
this function picks up twsk in any dying netns, not just the one passed
in via exit_batch list.
This means that the error unwind of setup_net() can "steal" and destroy
timewait sockets belonging to the exiting netns.
This allows the netns exit worker to proceed to call
WARN_ON_ONCE(!refcount_dec_and_test(&net->ipv4.tcp_death_row.tw_refcount));
without the expected 1 -> 0 transition, which then splats.
At same time, error unwind path that is also running inet_twsk_purge()
will splat as well:
WARNING: .. at lib/refcount.c:31 refcount_warn_saturate+0x1ed/0x210
...
refcount_dec include/linux/refcount.h:351 [inline]
inet_twsk_kill+0x758/0x9c0 net/ipv4/inet_timewait_sock.c:70
inet_twsk_deschedule_put net/ipv4/inet_timewait_sock.c:221
inet_twsk_purge+0x725/0x890 net/ipv4/inet_timewait_sock.c:304
tcp_sk_exit_batch+0x1c/0x170 net/ipv4/tcp_ipv4.c:3522
ops_exit_list+0x128/0x180 net/core/net_namespace.c:178
setup_net+0x714/0xb40 net/core/net_namespace.c:375
copy_net_ns+0x2f0/0x670 net/core/net_namespace.c:508
create_new_namespaces+0x3ea/0xb10 kernel/nsproxy.c:110
... because refcount_dec() of tw_refcount unexpectedly dropped to 0.
This doesn't seem like an actual bug (no tw sockets got lost and I don't
see a use-after-free) but as erroneous trigger of debug check.
Add a mutex to force strict ordering: the task that calls tcp_twsk_purge()
blocks other task from doing final _dec_and_test before mutex-owner has
removed all tw sockets of dying netns.
Fixes: e9bd0cca09d1 ("tcp: Don't allocate tcp_death_row outside of struct netns_ipv4.")
Reported-by: syzbot+8ea26396ff85d23a8929@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/0000000000003a5292061f5e4e19@google.com/
Link: https://lore.kernel.org/netdev/20240812140104.GA21559@breakpoint.cc/
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20240812222857.29837-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
ipv4_tcp_sk is a per-CPU variable and relies on disabled BH for its
locking. Without per-CPU locking in local_bh_disable() on PREEMPT_RT
this data structure requires explicit locking.
Make a struct with a sock member (original ipv4_tcp_sk) and a
local_lock_t and use local_lock_nested_bh() for locking. This change
adds only lockdep coverage and does not alter the functional behaviour
for !PREEMPT_RT.
Cc: David Ahern <dsahern@kernel.org>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://patch.msgid.link/20240620132727.660738-7-bigeasy@linutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Replace kfree_skb_reason with sk_skb_reason_drop and pass the receiving
socket to the tracepoint.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/r/202406011539.jhwBd7DX-lkp@intel.com/
Signed-off-by: Yan Zhai <yan@cloudflare.com>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
It's possible to clean-up some ifdefs by hiding that
tcp_{md5,ao}_needed static branch is defined and compiled only
under related configs, since commit 4c8530dc7d7d ("net/tcp: Only produce
AO/MD5 logs if there are any keys").
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Dmitry Safonov <0x7f454c46@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The TCP timewait timer is proving to be problematic for setups where
scheduler CPU isolation is achieved at runtime via cpusets (as opposed to
statically via isolcpus=domains).
What happens there is a CPU goes through tcp_time_wait(), arming the
time_wait timer, then gets isolated. TCP_TIMEWAIT_LEN later, the timer
fires, causing interference for the now-isolated CPU. This is conceptually
similar to the issue described in commit e02b93124855 ("workqueue: Unbind
kworkers before sending them to exit()")
Move inet_twsk_schedule() to within inet_twsk_hashdance(), with the ehash
lock held. Expand the lock's critical section from inet_twsk_kill() to
inet_twsk_deschedule_put(), serializing the scheduling vs descheduling of
the timer. IOW, this prevents the following race:
tcp_time_wait()
inet_twsk_hashdance()
inet_twsk_deschedule_put()
del_timer_sync()
inet_twsk_schedule()
Thanks to Paolo Abeni for suggesting to leverage the ehash lock.
This also restores a comment from commit ec94c2696f0b ("tcp/dccp: avoid
one atomic operation for timewait hashdance") as inet_twsk_hashdance() had
a "Step 1" and "Step 3" comment, but the "Step 2" had gone missing.
inet_twsk_deschedule_put() now acquires the ehash spinlock to synchronize
with inet_twsk_hashdance_schedule().
To ease possible regression search, actual un-pin is done in next patch.
Link: https://lore.kernel.org/all/ZPhpfMjSiHVjQkTk@localhost.localdomain/
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Valentin Schneider <vschneid@redhat.com>
Co-developed-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Adding a sysctl knob to allow user to specify a default
rto_min at socket init time, other than using the hard
coded 200ms default rto_min.
Note that the rto_min route option has the highest precedence
for configuring this setting, followed by the TCP_BPF_RTO_MIN
socket option, followed by the tcp_rto_min_us sysctl.
Signed-off-by: Kevin Yang <yyd@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Tony Lu <tonylu@linux.alibaba.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
These fields can be read and written locklessly, add annotations
around these minor races.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
tcp_skb_can_collapse() checks for conditions which don't make
sense on input. Because of this we ended up sprinkling a few
pairs of mptcp_skb_can_collapse() and skb_cmp_decrypted() calls
on the input path. Group them in a new helper. This should make
it less likely that someone will check mptcp and not decrypted
or vice versa when adding new code.
This implicitly adds a decrypted check early in tcp_collapse().
AFAIU this will very slightly increase our ability to collapse
packets under memory pressure, not a real bug.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Cross-merge networking fixes after downstream PR.
Conflicts:
drivers/net/ethernet/ti/icssg/icssg_classifier.c
abd5576b9c57 ("net: ti: icssg-prueth: Add support for ICSSG switch firmware")
56a5cf538c3f ("net: ti: icssg-prueth: Fix start counter for ft1 filter")
https://lore.kernel.org/all/20240531123822.3bb7eadf@canb.auug.org.au/
No other adjacent changes.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
These functions have races when they:
1) Write sk->sk_err
2) call sk_error_report(sk)
3) call tcp_done(sk)
As described in prior patches in this series:
An smp_wmb() is missing.
We should call tcp_done() before sk_error_report(sk)
to have consistent tcp_poll() results on SMP hosts.
Use tcp_done_with_error() where we centralized the
correct sequence.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Link: https://lore.kernel.org/r/20240528125253.1966136-5-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Jason commit made checks against ACK sequence less strict
and can be exploited by attackers to establish spoofed flows
with less probes.
Innocent users might use tcp_rmem[1] == 1,000,000,000,
or something more reasonable.
An attacker can use a regular TCP connection to learn the server
initial tp->rcv_wnd, and use it to optimize the attack.
If we make sure that only the announced window (smaller than 65535)
is used for ACK validation, we force an attacker to use
65537 packets to complete the 3WHS (assuming server ISN is unknown)
Fixes: 378979e94e95 ("tcp: remove 64 KByte limit for initial tp->rcv_wnd value")
Link: https://datatracker.ietf.org/meeting/119/materials/slides-119-tcpm-ghost-acks-00
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Jason Xing <kerneljasonxing@gmail.com>
Link: https://lore.kernel.org/r/20240523130528.60376-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
