We currently have one tcp bind table (bhash) which hashes by port
number only. In the socket bind path, we check for bind conflicts by
traversing the specified port's inet_bind2_bucket while holding the
bucket'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, checking for a bind conflict 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 proposes adding a second bind table, bhash2, that hashes by
port and ip address. Searching the bhash2 table leads to significantly
faster conflict resolution and less time holding the spinlock.
Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Acked-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
UDP sendmsg() is lockless, and reads sk->sk_bound_dev_if while
this field can be changed by another thread.
Adds minimal annotations to avoid KCSAN splats for UDP.
Following patches will add more annotations to potential lockless readers.
BUG: KCSAN: data-race in __ip6_datagram_connect / udpv6_sendmsg
write to 0xffff888136d47a94 of 4 bytes by task 7681 on cpu 0:
__ip6_datagram_connect+0x6e2/0x930 net/ipv6/datagram.c:221
ip6_datagram_connect+0x2a/0x40 net/ipv6/datagram.c:272
inet_dgram_connect+0x107/0x190 net/ipv4/af_inet.c:576
__sys_connect_file net/socket.c:1900 [inline]
__sys_connect+0x197/0x1b0 net/socket.c:1917
__do_sys_connect net/socket.c:1927 [inline]
__se_sys_connect net/socket.c:1924 [inline]
__x64_sys_connect+0x3d/0x50 net/socket.c:1924
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x2b/0x50 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
read to 0xffff888136d47a94 of 4 bytes by task 7670 on cpu 1:
udpv6_sendmsg+0xc60/0x16e0 net/ipv6/udp.c:1436
inet6_sendmsg+0x5f/0x80 net/ipv6/af_inet6.c:652
sock_sendmsg_nosec net/socket.c:705 [inline]
sock_sendmsg net/socket.c:725 [inline]
____sys_sendmsg+0x39a/0x510 net/socket.c:2413
___sys_sendmsg net/socket.c:2467 [inline]
__sys_sendmmsg+0x267/0x4c0 net/socket.c:2553
__do_sys_sendmmsg net/socket.c:2582 [inline]
__se_sys_sendmmsg net/socket.c:2579 [inline]
__x64_sys_sendmmsg+0x53/0x60 net/socket.c:2579
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x2b/0x50 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
value changed: 0x00000000 -> 0xffffff9b
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 7670 Comm: syz-executor.3 Tainted: G W 5.18.0-rc1-syzkaller-dirty #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
I chose to not add Fixes: tag because race has minor consequences
and stable teams busy enough.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
INET_MATCH() runs without holding a lock on the socket.
We probably need to annotate most reads.
This patch makes INET_MATCH() an inline function
to ease our changes.
v2:
We remove the 32bit version of it, as modern compilers
should generate the same code really, no need to
try to be smarter.
Also make 'struct net *net' the first argument.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sock_alloc_send_skb() is simple and just proxying to another function,
so we can inline it and cut associated overhead.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adding a new socket option, SO_RCVMARK, to indicate that SO_MARK
should be included in the ancillary data returned by recvmsg().
Renamed the sock_recv_ts_and_drops() function to sock_recv_cmsgs().
Signed-off-by: Erin MacNeil <lnx.erin@gmail.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>
Link: https://lore.kernel.org/r/20220427200259.2564-1-lnx.erin@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Logic added in commit f35f821935 ("tcp: defer skb freeing after socket
lock is released") helped bulk TCP flows to move the cost of skbs
frees outside of critical section where socket lock was held.
But for RPC traffic, or hosts with RFS enabled, the solution is far from
being ideal.
For RPC traffic, recvmsg() has to return to user space right after
skb payload has been consumed, meaning that BH handler has no chance
to pick the skb before recvmsg() thread. This issue is more visible
with BIG TCP, as more RPC fit one skb.
For RFS, even if BH handler picks the skbs, they are still picked
from the cpu on which user thread is running.
Ideally, it is better to free the skbs (and associated page frags)
on the cpu that originally allocated them.
This patch removes the per socket anchor (sk->defer_list) and
instead uses a per-cpu list, which will hold more skbs per round.
This new per-cpu list is drained at the end of net_action_rx(),
after incoming packets have been processed, to lower latencies.
In normal conditions, skbs are added to the per-cpu list with
no further action. In the (unlikely) cases where the cpu does not
run net_action_rx() handler fast enough, we use an IPI to raise
NET_RX_SOFTIRQ on the remote cpu.
Also, we do not bother draining the per-cpu list from dev_cpu_dead()
This is because skbs in this list have no requirement on how fast
they should be freed.
Note that we can add in the future a small per-cpu cache
if we see any contention on sd->defer_lock.
Tested on a pair of hosts with 100Gbit NIC, RFS enabled,
and /proc/sys/net/ipv4/tcp_rmem[2] tuned to 16MB to work around
page recycling strategy used by NIC driver (its page pool capacity
being too small compared to number of skbs/pages held in sockets
receive queues)
Note that this tuning was only done to demonstrate worse
conditions for skb freeing for this particular test.
These conditions can happen in more general production workload.
10 runs of one TCP_STREAM flow
Before:
Average throughput: 49685 Mbit.
Kernel profiles on cpu running user thread recvmsg() show high cost for
skb freeing related functions (*)
57.81% [kernel] [k] copy_user_enhanced_fast_string
(*) 12.87% [kernel] [k] skb_release_data
(*) 4.25% [kernel] [k] __free_one_page
(*) 3.57% [kernel] [k] __list_del_entry_valid
1.85% [kernel] [k] __netif_receive_skb_core
1.60% [kernel] [k] __skb_datagram_iter
(*) 1.59% [kernel] [k] free_unref_page_commit
(*) 1.16% [kernel] [k] __slab_free
1.16% [kernel] [k] _copy_to_iter
(*) 1.01% [kernel] [k] kfree
(*) 0.88% [kernel] [k] free_unref_page
0.57% [kernel] [k] ip6_rcv_core
0.55% [kernel] [k] ip6t_do_table
0.54% [kernel] [k] flush_smp_call_function_queue
(*) 0.54% [kernel] [k] free_pcppages_bulk
0.51% [kernel] [k] llist_reverse_order
0.38% [kernel] [k] process_backlog
(*) 0.38% [kernel] [k] free_pcp_prepare
0.37% [kernel] [k] tcp_recvmsg_locked
(*) 0.37% [kernel] [k] __list_add_valid
0.34% [kernel] [k] sock_rfree
0.34% [kernel] [k] _raw_spin_lock_irq
(*) 0.33% [kernel] [k] __page_cache_release
0.33% [kernel] [k] tcp_v6_rcv
(*) 0.33% [kernel] [k] __put_page
(*) 0.29% [kernel] [k] __mod_zone_page_state
0.27% [kernel] [k] _raw_spin_lock
After patch:
Average throughput: 73076 Mbit.
Kernel profiles on cpu running user thread recvmsg() looks better:
81.35% [kernel] [k] copy_user_enhanced_fast_string
1.95% [kernel] [k] _copy_to_iter
1.95% [kernel] [k] __skb_datagram_iter
1.27% [kernel] [k] __netif_receive_skb_core
1.03% [kernel] [k] ip6t_do_table
0.60% [kernel] [k] sock_rfree
0.50% [kernel] [k] tcp_v6_rcv
0.47% [kernel] [k] ip6_rcv_core
0.45% [kernel] [k] read_tsc
0.44% [kernel] [k] _raw_spin_lock_irqsave
0.37% [kernel] [k] _raw_spin_lock
0.37% [kernel] [k] native_irq_return_iret
0.33% [kernel] [k] __inet6_lookup_established
0.31% [kernel] [k] ip6_protocol_deliver_rcu
0.29% [kernel] [k] tcp_rcv_established
0.29% [kernel] [k] llist_reverse_order
v2: kdoc issue (kernel bots)
do not defer if (alloc_cpu == smp_processor_id()) (Paolo)
replace the sk_buff_head with a single-linked list (Jakub)
add a READ_ONCE()/WRITE_ONCE() for the lockless read of sd->defer_list
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Link: https://lore.kernel.org/r/20220422201237.416238-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The internal recvmsg() functions have two parameters 'flags' and 'noblock'
that were merged inside skb_recv_datagram(). As a follow up patch to commit
f4b41f062c ("net: remove noblock parameter from skb_recv_datagram()")
this patch removes the separate 'noblock' parameter for recvmsg().
Analogue to the referenced patch for skb_recv_datagram() the 'flags' and
'noblock' parameters are unnecessarily split up with e.g.
err = sk->sk_prot->recvmsg(sk, msg, size, flags & MSG_DONTWAIT,
flags & ~MSG_DONTWAIT, &addr_len);
or in
err = INDIRECT_CALL_2(sk->sk_prot->recvmsg, tcp_recvmsg, udp_recvmsg,
sk, msg, size, flags & MSG_DONTWAIT,
flags & ~MSG_DONTWAIT, &addr_len);
instead of simply using only flags all the time and check for MSG_DONTWAIT
where needed (to preserve for the formerly separated no(n)block condition).
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
Link: https://lore.kernel.org/r/20220411124955.154876-1-socketcan@hartkopp.net
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
In order to report the reasons of skb drops in 'sock_queue_rcv_skb()',
introduce the function 'sock_queue_rcv_skb_reason()'.
As the return value of 'sock_queue_rcv_skb()' is used as the error code,
we can't make it as drop reason and have to pass extra output argument.
'sock_queue_rcv_skb()' is used in many places, so we can't change it
directly.
Introduce the new function 'sock_queue_rcv_skb_reason()' and make
'sock_queue_rcv_skb()' an inline call to it.
Reviewed-by: Hao Peng <flyingpeng@tencent.com>
Reviewed-by: Jiang Biao <benbjiang@tencent.com>
Signed-off-by: Menglong Dong <imagedong@tencent.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
UDP sendmsg() can be lockless, this is causing all kinds
of data races.
This patch converts sk->sk_tskey to remove one of these races.
BUG: KCSAN: data-race in __ip_append_data / __ip_append_data
read to 0xffff8881035d4b6c of 4 bytes by task 8877 on cpu 1:
__ip_append_data+0x1c1/0x1de0 net/ipv4/ip_output.c:994
ip_make_skb+0x13f/0x2d0 net/ipv4/ip_output.c:1636
udp_sendmsg+0x12bd/0x14c0 net/ipv4/udp.c:1249
inet_sendmsg+0x5f/0x80 net/ipv4/af_inet.c:819
sock_sendmsg_nosec net/socket.c:705 [inline]
sock_sendmsg net/socket.c:725 [inline]
____sys_sendmsg+0x39a/0x510 net/socket.c:2413
___sys_sendmsg net/socket.c:2467 [inline]
__sys_sendmmsg+0x267/0x4c0 net/socket.c:2553
__do_sys_sendmmsg net/socket.c:2582 [inline]
__se_sys_sendmmsg net/socket.c:2579 [inline]
__x64_sys_sendmmsg+0x53/0x60 net/socket.c:2579
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x44/0xd0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
write to 0xffff8881035d4b6c of 4 bytes by task 8880 on cpu 0:
__ip_append_data+0x1d8/0x1de0 net/ipv4/ip_output.c:994
ip_make_skb+0x13f/0x2d0 net/ipv4/ip_output.c:1636
udp_sendmsg+0x12bd/0x14c0 net/ipv4/udp.c:1249
inet_sendmsg+0x5f/0x80 net/ipv4/af_inet.c:819
sock_sendmsg_nosec net/socket.c:705 [inline]
sock_sendmsg net/socket.c:725 [inline]
____sys_sendmsg+0x39a/0x510 net/socket.c:2413
___sys_sendmsg net/socket.c:2467 [inline]
__sys_sendmmsg+0x267/0x4c0 net/socket.c:2553
__do_sys_sendmmsg net/socket.c:2582 [inline]
__se_sys_sendmmsg net/socket.c:2579 [inline]
__x64_sys_sendmmsg+0x53/0x60 net/socket.c:2579
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x44/0xd0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x44/0xae
value changed: 0x0000054d -> 0x0000054e
Reported by Kernel Concurrency Sanitizer on:
CPU: 0 PID: 8880 Comm: syz-executor.5 Not tainted 5.17.0-rc2-syzkaller-00167-gdcb85f85fa6f-dirty #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Fixes: 09c2d251b7 ("net-timestamp: add key to disambiguate concurrent datagrams")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Willem de Bruijn <willemb@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add the SO_TXREHASH socket option to control hash rethink behavior per socket.
When default mode is set, sockets disable rehash at initialization and use
sysctl option when entering listen state. setsockopt() overrides default
behavior.
Signed-off-by: Akhmat Karakotov <hmukos@yandex-team.ru>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add a per ns sysctl that controls the txhash rethink behavior:
net.core.txrehash. When enabled, the same behavior is retained,
when disabled, rethink is not performed. Sysctl is enabled by default.
Signed-off-by: Akhmat Karakotov <hmukos@yandex-team.ru>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The return value of BPF_CGROUP_RUN_PROG_INET{4,6}_POST_BIND() in
__inet_bind() is not handled properly. While the return value
is non-zero, it will set inet_saddr and inet_rcv_saddr to 0 and
exit:
err = BPF_CGROUP_RUN_PROG_INET4_POST_BIND(sk);
if (err) {
inet->inet_saddr = inet->inet_rcv_saddr = 0;
goto out_release_sock;
}
Let's take UDP for example and see what will happen. For UDP
socket, it will be added to 'udp_prot.h.udp_table->hash' and
'udp_prot.h.udp_table->hash2' after the sk->sk_prot->get_port()
called success. If 'inet->inet_rcv_saddr' is specified here,
then 'sk' will be in the 'hslot2' of 'hash2' that it don't belong
to (because inet_saddr is changed to 0), and UDP packet received
will not be passed to this sock. If 'inet->inet_rcv_saddr' is not
specified here, the sock will work fine, as it can receive packet
properly, which is wired, as the 'bind()' is already failed.
To undo the get_port() operation, introduce the 'put_port' field
for 'struct proto'. For TCP proto, it is inet_put_port(); For UDP
proto, it is udp_lib_unhash(); For icmp proto, it is
ping_unhash().
Therefore, after sys_bind() fail caused by
BPF_CGROUP_RUN_PROG_INET4_POST_BIND(), it will be unbinded, which
means that it can try to be binded to another port.
Signed-off-by: Menglong Dong <imagedong@tencent.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20220106132022.3470772-2-imagedong@tencent.com
Alexei Starovoitov says:
====================
pull-request: bpf-next 2021-12-30
The following pull-request contains BPF updates for your *net-next* tree.
We've added 72 non-merge commits during the last 20 day(s) which contain
a total of 223 files changed, 3510 insertions(+), 1591 deletions(-).
The main changes are:
1) Automatic setrlimit in libbpf when bpf is memcg's in the kernel, from Andrii.
2) Beautify and de-verbose verifier logs, from Christy.
3) Composable verifier types, from Hao.
4) bpf_strncmp helper, from Hou.
5) bpf.h header dependency cleanup, from Jakub.
6) get_func_[arg|ret|arg_cnt] helpers, from Jiri.
7) Sleepable local storage, from KP.
8) Extend kfunc with PTR_TO_CTX, PTR_TO_MEM argument support, from Kumar.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
sock.h is pretty heavily used (5k objects rebuilt on x86 after
it's touched). We can drop the include of filter.h from it and
add a forward declaration of struct sk_filter instead.
This decreases the number of rebuilt objects when bpf.h
is touched from ~5k to ~1k.
There's a lot of missing includes this was masking. Primarily
in networking tho, this time.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Acked-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/bpf/20211229004913.513372-1-kuba@kernel.org
This patch moves sock_release_ownership() down in include/net/sock.h and
replaces some sk_lock.owned tests with sock_owned_by_user_nocheck().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Link: https://lore.kernel.org/r/20211208062158.54132-1-kuniyu@amazon.co.jp
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Steffen reported a TCP stream corruption for HTTP requests
served by the apache web-server using a cifs mount-point
and memory mapping the relevant file.
The root cause is quite similar to the one addressed by
commit 20eb4f29b6 ("net: fix sk_page_frag() recursion from
memory reclaim"). Here the nested access to the task page frag
is caused by a page fault on the (mmapped) user-space memory
buffer coming from the cifs file.
The page fault handler performs an smb transaction on a different
socket, inside the same process context. Since sk->sk_allaction
for such socket does not prevent the usage for the task_frag,
the nested allocation modify "under the hood" the page frag
in use by the outer sendmsg call, corrupting the stream.
The overall relevant stack trace looks like the following:
httpd 78268 [001] 3461630.850950: probe:tcp_sendmsg_locked:
ffffffff91461d91 tcp_sendmsg_locked+0x1
ffffffff91462b57 tcp_sendmsg+0x27
ffffffff9139814e sock_sendmsg+0x3e
ffffffffc06dfe1d smb_send_kvec+0x28
[...]
ffffffffc06cfaf8 cifs_readpages+0x213
ffffffff90e83c4b read_pages+0x6b
ffffffff90e83f31 __do_page_cache_readahead+0x1c1
ffffffff90e79e98 filemap_fault+0x788
ffffffff90eb0458 __do_fault+0x38
ffffffff90eb5280 do_fault+0x1a0
ffffffff90eb7c84 __handle_mm_fault+0x4d4
ffffffff90eb8093 handle_mm_fault+0xc3
ffffffff90c74f6d __do_page_fault+0x1ed
ffffffff90c75277 do_page_fault+0x37
ffffffff9160111e page_fault+0x1e
ffffffff9109e7b5 copyin+0x25
ffffffff9109eb40 _copy_from_iter_full+0xe0
ffffffff91462370 tcp_sendmsg_locked+0x5e0
ffffffff91462370 tcp_sendmsg_locked+0x5e0
ffffffff91462b57 tcp_sendmsg+0x27
ffffffff9139815c sock_sendmsg+0x4c
ffffffff913981f7 sock_write_iter+0x97
ffffffff90f2cc56 do_iter_readv_writev+0x156
ffffffff90f2dff0 do_iter_write+0x80
ffffffff90f2e1c3 vfs_writev+0xa3
ffffffff90f2e27c do_writev+0x5c
ffffffff90c042bb do_syscall_64+0x5b
ffffffff916000ad entry_SYSCALL_64_after_hwframe+0x65
The cifs filesystem rightfully sets sk_allocations to GFP_NOFS,
we can avoid the nesting using the sk page frag for allocation
lacking the __GFP_FS flag. Do not define an additional mm-helper
for that, as this is strictly tied to the sk page frag usage.
v1 -> v2:
- use a stricted sk_page_frag() check instead of reordering the
code (Eric)
Reported-by: Steffen Froemer <sfroemer@redhat.com>
Fixes: 5640f76858 ("net: use a per task frag allocator")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add neigh_confirm() for the confirmed member in struct neighbour,
it can be called as an independent unit by other functions.
Signed-off-by: Yajun Deng <yajun.deng@linux.dev>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add missing entries to fix these "make htmldocs" warnings.
./include/linux/skbuff.h:953: warning: Function parameter or member 'll_node' not described in 'sk_buff'
./include/net/sock.h:540: warning: Function parameter or member 'defer_list' not described in 'sock'
Fixes: f35f821935 ("tcp: defer skb freeing after socket lock is released")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This is distracting really, let's make this simpler,
because many callers had to take care of this
by themselves, even if on x86 this adds more
code than really needed.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net->core.sock_inuse is a per cpu variable (int),
while net->core.prot_inuse is another per cpu variable
of 64 integers.
per cpu allocator tend to place them in very different places.
Grouping them together makes sense, since it makes
updates potentially faster, if hitting the same
cache line.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MPTCP hard codes it, let us instead provide this helper.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sock_prot_inuse_add() is very small, we can inline it.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk_rx_dst/sk_rx_dst_ifindex/sk_rx_dst_cookie are read in early demux,
and currently spans two cache lines.
Moving them close to sk_refcnt makes more sense, as only one cache
line is needed.
New layout for this hot cache line is :
struct sock {
struct sock_common __sk_common; /* 0 0x88 */
/* --- cacheline 2 boundary (128 bytes) was 8 bytes ago --- */
struct dst_entry * sk_rx_dst; /* 0x88 0x8 */
int sk_rx_dst_ifindex; /* 0x90 0x4 */
u32 sk_rx_dst_cookie; /* 0x94 0x4 */
socket_lock_t sk_lock; /* 0x98 0x20 */
atomic_t sk_drops; /* 0xb8 0x4 */
int sk_rcvlowat; /* 0xbc 0x4 */
/* --- cacheline 3 boundary (192 bytes) --- */
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tcp recvmsg() (or rx zerocopy) spends a fair amount of time
freeing skbs after their payload has been consumed.
A typical ~64KB GRO packet has to release ~45 page
references, eventually going to page allocator
for each of them.
Currently, this freeing is performed while socket lock
is held, meaning that there is a high chance that
BH handler has to queue incoming packets to tcp socket backlog.
This can cause additional latencies, because the user
thread has to process the backlog at release_sock() time,
and while doing so, additional frames can be added
by BH handler.
This patch adds logic to defer these frees after socket
lock is released, or directly from BH handler if possible.
Being able to free these skbs from BH handler helps a lot,
because this avoids the usual alloc/free assymetry,
when BH handler and user thread do not run on same cpu or
NUMA node.
One cpu can now be fully utilized for the kernel->user copy,
and another cpu is handling BH processing and skb/page
allocs/frees (assuming RFS is not forcing use of a single CPU)
Tested:
100Gbit NIC
Max throughput for one TCP_STREAM flow, over 10 runs
MTU : 1500
Before: 55 Gbit
After: 66 Gbit
MTU : 4096+(headers)
Before: 82 Gbit
After: 95 Gbit
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use INDIRECT_CALL_INET() to avoid an indirect call
when/if CONFIG_RETPOLINE=y
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
(struct proto)->sk_forward_alloc is currently only used by MPTCP.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move sk_bind_phc next to sk_peer_lock to fill a hole.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Instead of using a full netdev_features_t, we can use a single bit,
as sk_route_nocaps is only used to remove NETIF_F_GSO_MASK from
sk->sk_route_cap.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We were only using one bit, and we can replace it by sk_is_tcp()
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move sk_is_tcp() to include/net/sock.h and use it where we can.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk_wmem_free_skb() is only used by TCP.
Rename it to make this clear, and move its declaration to
include/net/tcp.h
Signed-off-by: Talal Ahmad <talalahmad@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Acked-by: Arjun Roy <arjunroy@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We now have INDIRECT_CALL_INET_1() macro, no need to use #ifdef CONFIG_INET
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A later patch will change the MPTCP memory accounting schema
in such a way that MPTCP sockets will encode the total amount of
forward allocated memory in two separate fields (one for tx and
one for rx).
MPTCP sockets will use their own helper to provide the accurate
amount of fwd allocated memory.
To allow the above, this patch adds a new, optional, sk method to
fetch the fwd memory, wrap the call in a new helper and use it
where it is appropriate.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
A following patch is going to implement a similar reclaim schema
for the MPTCP protocol, with different locking.
Let's define a couple of macros for the used thresholds, so
that the latter code will be more easily maintainable.
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The proto ops ->stream_memory_read() is currently only used
by TCP to check whether psock queue is empty or not. We need
to rename it before reusing it for non-TCP protocols, and
adjust the exsiting users accordingly.
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20211008203306.37525-2-xiyou.wangcong@gmail.com
sk_stream_alloc_skb() is only used by TCP.
Rename it to make this clear, and move its declaration
to include/net/tcp.h
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>
sk->sk_rx_queue_mapping can be modified locklessly,
add a couple of READ_ONCE()/WRITE_ONCE() to document this fact.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
sk_rx_queue_mapping is located in a cache line that should be kept read mostly.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Increase cache locality by moving rx_dst_coookie next to sk->sk_rx_dst
This removes one or two cache line misses in IPv6 early demux (TCP/UDP)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Increase cache locality by moving rx_dst_ifindex next to sk->sk_rx_dst
This is part of an effort to reduce cache line misses in TCP fast path.
This removes one cache line miss in early demux.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Use of percpu_counter structure to track count of orphaned
sockets is causing problems on modern hosts with 256 cpus
or more.
Stefan Bach reported a serious spinlock contention in real workloads,
that I was able to reproduce with a netfilter rule dropping
incoming FIN packets.
53.56% server [kernel.kallsyms] [k] queued_spin_lock_slowpath
|
---queued_spin_lock_slowpath
|
--53.51%--_raw_spin_lock_irqsave
|
--53.51%--__percpu_counter_sum
tcp_check_oom
|
|--39.03%--__tcp_close
| tcp_close
| inet_release
| inet6_release
| sock_close
| __fput
| ____fput
| task_work_run
| exit_to_usermode_loop
| do_syscall_64
| entry_SYSCALL_64_after_hwframe
| __GI___libc_close
|
--14.48%--tcp_out_of_resources
tcp_write_timeout
tcp_retransmit_timer
tcp_write_timer_handler
tcp_write_timer
call_timer_fn
expire_timers
__run_timers
run_timer_softirq
__softirqentry_text_start
As explained in commit cf86a086a1 ("net/dst: use a smaller percpu_counter
batch for dst entries accounting"), default batch size is too big
for the default value of tcp_max_orphans (262144).
But even if we reduce batch sizes, there would still be cases
where the estimated count of orphans is beyond the limit,
and where tcp_too_many_orphans() has to call the expensive
percpu_counter_sum_positive().
One solution is to use plain per-cpu counters, and have
a timer to periodically refresh this cache.
Updating this cache every 100ms seems about right, tcp pressure
state is not radically changing over shorter periods.
percpu_counter was nice 15 years ago while hosts had less
than 16 cpus, not anymore by current standards.
v2: Fix the build issue for CONFIG_CRYPTO_DEV_CHELSIO_TLS=m,
reported by kernel test robot <lkp@intel.com>
Remove unused socket argument from tcp_too_many_orphans()
Fixes: dd24c00191 ("net: Use a percpu_counter for orphan_count")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Stefan Bach <sfb@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Reuse the timeval compat code from core/sock to handle 32-bit and
64-bit timeval structures. Also introduce a new socket option define
to allow using y2038 safe timeval under 32-bit.
The existing behavior of sock_set_timeout and vsock's timeout setter
differ when the time value is out of bounds. vsocks current behavior
is retained at the expense of not being able to share the full
implementation.
This allows the LTP test vsock01 to pass under 32-bit compat mode.
Fixes: fe0c72f3db ("socket: move compat timeout handling into sock.c")
Signed-off-by: Richard Palethorpe <rpalethorpe@suse.com>
Cc: Richard Palethorpe <rpalethorpe@richiejp.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
