[ Upstream commit 1eeb5043573981f3a1278876515851b7f6b1df1b ]
We lost ability to unload ipv6 module a long time ago.
Instead of calling expensive inet_twsk_purge() twice,
we can handle all families in one round.
Also remove an extra line added in my prior patch,
per Kuniyuki Iwashima feedback.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/netdev/20240327192934.6843-1-kuniyu@amazon.com/
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20240329153203.345203-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 565d121b6998 ("tcp: prevent concurrent execution of tcp_sk_exit_batch")
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit ff46e3b4421923937b7f6e44ffcd3549a074f321 ]
When bonding is configured in BOND_MODE_BROADCAST mode, if two identical
SYN packets are received at the same time and processed on different CPUs,
it can potentially create the same sk (sock) but two different reqsk
(request_sock) in tcp_conn_request().
These two different reqsk will respond with two SYNACK packets, and since
the generation of the seq (ISN) incorporates a timestamp, the final two
SYNACK packets will have different seq values.
The consequence is that when the Client receives and replies with an ACK
to the earlier SYNACK packet, we will reset(RST) it.
========================================================================
This behavior is consistently reproducible in my local setup,
which comprises:
| NETA1 ------ NETB1 |
PC_A --- bond --- | | --- bond --- PC_B
| NETA2 ------ NETB2 |
- PC_A is the Server and has two network cards, NETA1 and NETA2. I have
bonded these two cards using BOND_MODE_BROADCAST mode and configured
them to be handled by different CPU.
- PC_B is the Client, also equipped with two network cards, NETB1 and
NETB2, which are also bonded and configured in BOND_MODE_BROADCAST mode.
If the client attempts a TCP connection to the server, it might encounter
a failure. Capturing packets from the server side reveals:
10.10.10.10.45182 > localhost: Flags [S], seq 320236027,
10.10.10.10.45182 > localhost: Flags [S], seq 320236027,
localhost > 10.10.10.10.45182: Flags [S.], seq 2967855116,
localhost > 10.10.10.10.45182: Flags [S.], seq 2967855123, <==
10.10.10.10.45182 > localhost: Flags [.], ack 4294967290,
10.10.10.10.45182 > localhost: Flags [.], ack 4294967290,
localhost > 10.10.10.10.45182: Flags [R], seq 2967855117, <==
localhost > 10.10.10.10.45182: Flags [R], seq 2967855117,
Two SYNACKs with different seq numbers are sent by localhost,
resulting in an anomaly.
========================================================================
The attempted solution is as follows:
Add a return value to inet_csk_reqsk_queue_hash_add() to confirm if the
ehash insertion is successful (Up to now, the reason for unsuccessful
insertion is that a reqsk for the same connection has already been
inserted). If the insertion fails, release the reqsk.
Due to the refcnt, Kuniyuki suggests also adding a return value check
for the DCCP module; if ehash insertion fails, indicating a successful
insertion of the same connection, simply release the reqsk as well.
Simultaneously, In the reqsk_queue_hash_req(), the start of the
req->rsk_timer is adjusted to be after successful insertion.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: luoxuanqiang <luoxuanqiang@kylinos.cn>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20240621013929.1386815-1-luoxuanqiang@kylinos.cn
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 23be1e0e2a83a8543214d2599a31d9a2185a796b ]
Initially, commit 4237c75c0a ("[MLSXFRM]: Auto-labeling of child
sockets") introduced security_inet_conn_request() in some functions
where reqsk is allocated. The hook is added just after the allocation,
so reqsk's IPv6 remote address was not initialised then.
However, SELinux/Smack started to read it in netlbl_req_setattr()
after commit e1adea9270 ("calipso: Allow request sockets to be
relabelled by the lsm.").
Commit 284904aa79 ("lsm: Relocate the IPv4 security_inet_conn_request()
hooks") fixed that kind of issue only in TCPv4 because IPv6 labeling was
not supported at that time. Finally, the same issue was introduced again
in IPv6.
Let's apply the same fix on DCCPv6 and TCPv6.
Fixes: e1adea9270 ("calipso: Allow request sockets to be relabelled by the lsm.")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit fa2df45af13091f76b89adb84a28f13818d5d631 ]
Initially, commit 4237c75c0a ("[MLSXFRM]: Auto-labeling of child
sockets") introduced security_inet_conn_request() in some functions
where reqsk is allocated. The hook is added just after the allocation,
so reqsk's IPv4 remote address was not initialised then.
However, SELinux/Smack started to read it in netlbl_req_setattr()
after the cited commits.
This bug was partially fixed by commit 284904aa79 ("lsm: Relocate
the IPv4 security_inet_conn_request() hooks").
This patch fixes the last bug in DCCPv4.
Fixes: 389fb800ac ("netlabel: Label incoming TCP connections correctly in SELinux")
Fixes: 07feee8f81 ("netlabel: Cleanup the Smack/NetLabel code to fix incoming TCP connections")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
There was a previous attempt to fix an out-of-bounds access in the DCCP
error handlers, but that fix assumed that the error handlers only want
to access the first 8 bytes of the DCCP header. Actually, they also look
at the DCCP sequence number, which is stored beyond 8 bytes, so an
explicit pskb_may_pull() is required.
Fixes: 6706a97fec ("dccp: fix out of bound access in dccp_v4_err()")
Fixes: 1aa9d1a0e7 ("ipv6: dccp: fix out of bound access in dccp_v6_err()")
Cc: stable@vger.kernel.org
Signed-off-by: Jann Horn <jannh@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
We changed tcp_poll() over time, bug never updated dccp.
Note that we also could remove dccp instead of maintaining it.
Fixes: 7c657876b6 ("[DCCP]: Initial implementation")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230818015820.2701595-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
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>
dccp_sendmsg() reads dp->dccps_mss_cache before locking the socket.
Same thing in do_dccp_getsockopt().
Add READ_ONCE()/WRITE_ONCE() annotations,
and change dccp_sendmsg() to check again dccps_mss_cache
after socket is locked.
Fixes: 7c657876b6 ("[DCCP]: Initial implementation")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230803163021.2958262-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
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>
IPv6 inet sockets are supposed to have a "struct ipv6_pinfo"
field at the end of their definition, so that inet6_sk_generic()
can derive from socket size the offset of the "struct ipv6_pinfo".
This is very fragile, and prevents adding bigger alignment
in sockets, because inet6_sk_generic() does not work
if the compiler adds padding after the ipv6_pinfo component.
We are currently working on a patch series to reorganize
TCP structures for better data locality and found issues
similar to the one fixed in commit f5d547676c
("tcp: fix tcp_inet6_sk() for 32bit kernels")
Alternative would be to force an alignment on "struct ipv6_pinfo",
greater or equal to __alignof__(any ipv6 sock) to ensure there is
no padding. This does not look great.
v2: fix typo in mptcp_proto_v6_init() (Paolo)
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Chao Wu <wwchao@google.com>
Cc: Wei Wang <weiwan@google.com>
Cc: Coco Li <lixiaoyan@google.com>
Cc: YiFei Zhu <zhuyifei@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
There's no reason for setting the RTO_ONLINK flag in ->flowi4_tos as
RT_CONN_FLAGS() does. We can easily set ->flowi4_scope properly
instead. This makes the code more explicit and will allow to convert
->flowi4_tos to dscp_t in the future.
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Most of the ioctls to net protocols operates directly on userspace
argument (arg). Usually doing get_user()/put_user() directly in the
ioctl callback. This is not flexible, because it is hard to reuse these
functions without passing userspace buffers.
Change the "struct proto" ioctls to avoid touching userspace memory and
operate on kernel buffers, i.e., all protocol's ioctl callbacks is
adapted to operate on a kernel memory other than on userspace (so, no
more {put,get}_user() and friends being called in the ioctl callback).
This changes the "struct proto" ioctl format in the following way:
int (*ioctl)(struct sock *sk, int cmd,
- unsigned long arg);
+ int *karg);
(Important to say that this patch does not touch the "struct proto_ops"
protocols)
So, the "karg" argument, which is passed to the ioctl callback, is a
pointer allocated to kernel space memory (inside a function wrapper).
This buffer (karg) may contain input argument (copied from userspace in
a prep function) and it might return a value/buffer, which is copied
back to userspace if necessary. There is not one-size-fits-all format
(that is I am using 'may' above), but basically, there are three type of
ioctls:
1) Do not read from userspace, returns a result to userspace
2) Read an input parameter from userspace, and does not return anything
to userspace
3) Read an input from userspace, and return a buffer to userspace.
The default case (1) (where no input parameter is given, and an "int" is
returned to userspace) encompasses more than 90% of the cases, but there
are two other exceptions. Here is a list of exceptions:
* Protocol RAW:
* cmd = SIOCGETVIFCNT:
* input and output = struct sioc_vif_req
* cmd = SIOCGETSGCNT
* input and output = struct sioc_sg_req
* Explanation: for the SIOCGETVIFCNT case, userspace passes the input
argument, which is struct sioc_vif_req. Then the callback populates
the struct, which is copied back to userspace.
* Protocol RAW6:
* cmd = SIOCGETMIFCNT_IN6
* input and output = struct sioc_mif_req6
* cmd = SIOCGETSGCNT_IN6
* input and output = struct sioc_sg_req6
* Protocol PHONET:
* cmd == SIOCPNADDRESOURCE | SIOCPNDELRESOURCE
* input int (4 bytes)
* Nothing is copied back to userspace.
For the exception cases, functions sock_sk_ioctl_inout() will
copy the userspace input, and copy it back to kernel space.
The wrapper that prepare the buffer and put the buffer back to user is
sk_ioctl(), so, instead of calling sk->sk_prot->ioctl(), the callee now
calls sk_ioctl(), which will handle all cases.
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230609152800.830401-1-leitao@debian.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
DCCP was marked as Orphan in the MAINTAINERS entry 2 years ago in commit
054c4610bd ("MAINTAINERS: dccp: move Gerrit Renker to CREDITS"). It says
we haven't heard from the maintainer for five years, so DCCP is not well
maintained for 7 years now.
Recently DCCP only receives updates for bugs, and major distros disable it
by default.
Removing DCCP would allow for better organisation of TCP fields to reduce
the number of cache lines hit in the fast path.
Let's add a deprecation notice when DCCP socket is created and schedule its
removal to 2025.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Keep the conntrack reference until policy checks have been performed for
IPsec V6 NAT support, just like ipv4.
The reference needs to be dropped before a packet is
queued to avoid having the conntrack module unloadable.
Fixes: 58a317f106 ("netfilter: ipv6: add IPv6 NAT support")
Signed-off-by: Madhu Koriginja <madhu.koriginja@nxp.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
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>
Eric Dumazet pointed out [0] that when we call skb_set_owner_r()
for ipv6_pinfo.pktoptions, sk_rmem_schedule() has not been called,
resulting in a negative sk_forward_alloc.
We add a new helper which clones a skb and sets its owner only
when sk_rmem_schedule() succeeds.
Note that we move skb_set_owner_r() forward in (dccp|tcp)_v6_do_rcv()
because tcp_send_synack() can make sk_forward_alloc negative before
ipv6_opt_accepted() in the crossed SYN-ACK or self-connect() cases.
[0]: https://lore.kernel.org/netdev/CANn89iK9oc20Jdi_41jb9URdF210r7d1Y-+uypbMSbOfY6jqrg@mail.gmail.com/
Fixes: 323fbd0edf ("net: dccp: Add handling of IPV6_PKTOPTIONS to dccp_v6_do_rcv()")
Fixes: 3df80d9320 ("[DCCP]: Introduce DCCPv6")
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
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>
After commit d38afeec26 ("tcp/udp: Call inet6_destroy_sock()
in IPv6 sk->sk_destruct()."), we call inet6_destroy_sock() in
sk->sk_destruct() by setting inet6_sock_destruct() to it to make
sure we do not leak inet6-specific resources.
DCCP sets its own sk->sk_destruct() in the dccp_init_sock(), and
DCCPv6 socket shares it by calling the same init function via
dccp_v6_init_sock().
To call inet6_sock_destruct() from DCCPv6 sk->sk_destruct(), we
export it and set dccp_v6_sk_destruct() in the init function.
Signed-off-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>
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>
In the case of sk->dccps_qpolicy == DCCPQ_POLICY_PRIO, dccp_qpolicy_full
will drop a skb when qpolicy is full. And the lock in dccp_sendmsg is
released before sock_alloc_send_skb and then relocked after
sock_alloc_send_skb. The following conditions may lead dccp_qpolicy_push
to add skb to an already full sk_write_queue:
thread1--->lock
thread1--->dccp_qpolicy_full: queue is full. drop a skb
thread1--->unlock
thread2--->lock
thread2--->dccp_qpolicy_full: queue is not full. no need to drop.
thread2--->unlock
thread1--->lock
thread1--->dccp_qpolicy_push: add a skb. queue is full.
thread1--->unlock
thread2--->lock
thread2--->dccp_qpolicy_push: add a skb!
thread2--->unlock
Fix this by moving dccp_qpolicy_full.
Fixes: b1308dc015 ("[DCCP]: Set TX Queue Length Bounds via Sysctl")
Signed-off-by: Hangyu Hua <hbh25y@gmail.com>
Link: https://lore.kernel.org/r/20220729110027.40569-1-hbh25y@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
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>
When reading listener sk->sk_bound_dev_if locklessly,
we must use READ_ONCE().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commits:
0dad4087a8 ("tcp/dccp: get rid of inet_twsk_purge()")
d507204d3c ("tcp/dccp: add tw->tw_bslot")
As Leonard pointed out, a newly allocated netns can happen
to reuse a freed 'struct net'.
While TCP TW timers were covered by my patches, other things were not:
1) Lookups in rx path (INET_MATCH() and INET6_MATCH()), as they look
at 4-tuple plus the 'struct net' pointer.
2) /proc/net/tcp[6] and inet_diag, same reason.
3) hashinfo->bhash[], same reason.
Fixing all this seems risky, lets instead revert.
In the future, we might have a per netns tcp hash table, or
a per netns list of timewait sockets...
Fixes: 0dad4087a8 ("tcp/dccp: get rid of inet_twsk_purge()")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Leonard Crestez <cdleonard@gmail.com>
Tested-by: Leonard Crestez <cdleonard@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The listen sk is currently stored in two hash tables,
listening_hash (hashed by port) and lhash2 (hashed by port and address).
After commit 0ee58dad5b ("net: tcp6: prefer listeners bound to an address")
and commit d9fbc7f643 ("net: tcp: prefer listeners bound to an address"),
the TCP-SYN lookup fast path does not use listening_hash.
The commit 05c0b35709 ("tcp: seq_file: Replace listening_hash with lhash2")
also moved the seq_file (/proc/net/tcp) iteration usage from
listening_hash to lhash2.
There are still a few listening_hash usages left.
One of them is inet_reuseport_add_sock() which uses the listening_hash
to search a listen sk during the listen() system call. This turns
out to be very slow on use cases that listen on many different
VIPs at a popular port (e.g. 443). [ On top of the slowness in
adding to the tail in the IPv6 case ]. The latter patch has a
selftest to demonstrate this case.
This patch takes this chance to move all remaining listening_hash
usages to lhash2 and then retire listening_hash.
Since most changes need to be done together, it is hard to cut
the listening_hash to lhash2 switch into small patches. The
changes in this patch is highlighted here for the review
purpose.
1. Because of the listening_hash removal, lhash2 can use the
sk->sk_nulls_node instead of the icsk->icsk_listen_portaddr_node.
This will also keep the sk_unhashed() check to work as is
after stop adding sk to listening_hash.
The union is removed from inet_listen_hashbucket because
only nulls_head is needed.
2. icsk->icsk_listen_portaddr_node and its helpers are removed.
3. The current lhash2 users needs to iterate with sk_nulls_node
instead of icsk_listen_portaddr_node.
One case is in the inet[6]_lhash2_lookup().
Another case is the seq_file iterator in tcp_ipv4.c.
One thing to note is sk_nulls_next() is needed
because the old inet_lhash2_for_each_icsk_continue()
does a "next" first before iterating.
4. Move the remaining listening_hash usage to lhash2
inet_reuseport_add_sock() which this series is
trying to improve.
inet_diag.c and mptcp_diag.c are the final two
remaining use cases and is moved to lhash2 now also.
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Now that ip_rt_fix_tos() doesn't reset ->flowi4_scope unconditionally,
we don't have to rely on the RTO_ONLINK bit to properly set the scope
of a flowi4 structure. We can just set ->flowi4_scope explicitly and
avoid using RTO_ONLINK in ->flowi4_tos.
This patch converts callers of ip_route_connect(). Instead of setting
the tos parameter with RT_CONN_FLAGS(sk), as all callers do, we can:
1- Drop the tos parameter from ip_route_connect(): its value was
entirely based on sk, which is also passed as parameter.
2- Set ->flowi4_scope depending on the SOCK_LOCALROUTE socket option
instead of always initialising it with RT_SCOPE_UNIVERSE (let's
define ip_sock_rt_scope() for this purpose).
3- Avoid overloading ->flowi4_tos with RTO_ONLINK: since the scope is
now properly initialised, we don't need to tell ip_rt_fix_tos() to
adjust ->flowi4_scope for us. So let's define ip_sock_rt_tos(),
which is the same as RT_CONN_FLAGS() but without the RTO_ONLINK
bit overload.
Note:
In the original ip_route_connect() code, __ip_route_output_key()
might clear the RTO_ONLINK bit of fl4->flowi4_tos (because of
ip_rt_fix_tos()). Therefore flowi4_update_output() had to reuse the
original tos variable. Now that we don't set RTO_ONLINK any more,
this is not a problem and we can use fl4->flowi4_tos in
flowi4_update_output().
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since commit 9fe516ba3f ("inet: move ipv6only in sock_common"),
ipv6_only_sock() and __ipv6_only_sock() are the same macro. Let's
remove the one.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
Prior patches in the series made sure tw_timer_handler()
can be fired after netns has been dismantled/freed.
We no longer have to scan a potentially big TCP ehash
table at netns dismantle.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch inlines dccp_listen_start() and removes a stale comment in
inet_dccp_listen() so that it looks like inet_listen().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Reviewed-by: Richard Sailer <richard_siegfried@systemli.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The commit 1295e2cf30 ("inet: minor optimization for backlog setting in
listen(2)") added change so that sk_max_ack_backlog is initialised earlier
in inet_dccp_listen() and inet_listen(). Since then, we no longer use
backlog in inet_csk_listen_start(), so let's remove it.
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.co.jp>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Reviewed-by: Richard Sailer <richard_siegfried@systemli.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In preparation for FORTIFY_SOURCE performing compile-time and run-time
field bounds checking for memset(), avoid intentionally writing across
neighboring fields.
Use memset_startat() so memset() doesn't get confused about writing
beyond the destination member that is intended to be the starting point
of zeroing through the end of the struct.
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
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>
Commit 2677d20677 ("dccp: don't free ccid2_hc_tx_sock ...") fixed
a UAF but reintroduced CVE-2017-6074.
When the sock is cloned, two dccps_hc_tx_ccid will reference to the
same ccid. So one can free the ccid object twice from two socks after
cloning.
This issue was found by "Hadar Manor" as well and assigned with
CVE-2020-16119, which was fixed in Ubuntu's kernel. So here I port
the patch from Ubuntu to fix it.
The patch prevents cloned socks from referencing the same ccid.
Fixes: 2677d20677 ("dccp: don't free ccid2_hc_tx_sock ...")
Signed-off-by: Zhenpeng Lin <zplin@psu.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet