Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

fallback rx path is broken #200

Closed
pabeni opened this issue Jun 4, 2021 · 1 comment
Closed

fallback rx path is broken #200

pabeni opened this issue Jun 4, 2021 · 1 comment
Assignees
@pabeni
Labels
bug

Comments

@pabeni
Copy link

pabeni commented Jun 4, 2021

ok, it's almost as drastic as it sounds:

In fallback mode, tcp_rcv_established() we hit the fastpath/header prediction. Incoming data is usually enqueued via direct call to tcp_queue_rcv() and the mptcp_incoming_options() hook in tcp_data_queue() is usually not reached.

As a result, we could end-up delaying the writer wake-up or even stalling it completely.

That is somewhat perceivable even with self-tests: sometimes fallback self-tests are considerably slower than non fallback one.

Possible solutions:

  • explicitly disable tcp header prediction for fallback socket, e.g. setting a 0 'tp->tcp_header_len'
  • add more mptcp hooks in tcp()
  • ??? both the above are ugly!!!
@pabeni pabeni added the bug label Jun 4, 2021
@mjmartineau
Copy link
Member

If the maintainers won't accept one more unlikely() branch on the fast path (hey, md5 gets a call in tcp_v4_rcv()), I don't know what option we have other than turning off header prediction. Can anything be plugged in elsewhere on the rx path that could nudge the MPTCP code?

fengguang pushed a commit to 0day-ci/linux that referenced this issue Jun 17, 2021
@intel-lab-lkp
tcp: consistently disable header prediction for mptcp
f10c8c6 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: multipath-tcp/mptcp_net-next#200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
matttbe pushed a commit that referenced this issue Jun 24, 2021
@matttbe
tcp: consistently disable header prediction for mptcp
c33c021 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
jenkins-tessares pushed a commit that referenced this issue Jun 25, 2021
@jenkins-tessares
tcp: consistently disable header prediction for mptcp
2311eab 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
matttbe pushed a commit that referenced this issue Jun 25, 2021
@matttbe
tcp: consistently disable header prediction for mptcp
425c8ef 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
jenkins-tessares pushed a commit that referenced this issue Jun 26, 2021
@jenkins-tessares
tcp: consistently disable header prediction for mptcp
ea2802e 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
matttbe pushed a commit that referenced this issue Jun 28, 2021
@matttbe
tcp: consistently disable header prediction for mptcp
366ae1d 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
jenkins-tessares pushed a commit that referenced this issue Jun 29, 2021
@jenkins-tessares
tcp: consistently disable header prediction for mptcp
fd37081 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
jenkins-tessares pushed a commit that referenced this issue Jun 30, 2021
@jenkins-tessares
tcp: consistently disable header prediction for mptcp
ce39482 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
fengguang pushed a commit to 0day-ci/linux that referenced this issue Jun 30, 2021
@intel-lab-lkp
tcp: consistently disable header prediction for mptcp
f04a5f5 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when an MPTCP socket falls back to TCP, it can hit the TCP receive
fast-path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: multipath-tcp/mptcp_net-next#200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
jenkins-tessares pushed a commit that referenced this issue Jul 1, 2021
@jenkins-tessares
tcp: consistently disable header prediction for mptcp
fac3dbc 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
fengguang pushed a commit to 0day-ci/linux that referenced this issue Jul 1, 2021
@davem330
tcp: consistently disable header prediction for mptcp
71158bb 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when an MPTCP socket falls back to TCP, it can hit the TCP receive
fast-path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: multipath-tcp/mptcp_net-next#200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
jenkins-tessares pushed a commit that referenced this issue Jul 2, 2021
@jenkins-tessares
tcp: consistently disable header prediction for mptcp
debfc5f 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
matttbe pushed a commit that referenced this issue Jul 7, 2021
@matttbe
tcp: consistently disable header prediction for mptcp
cdf2c7b 
The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when and MPTCP socket falls back to TCP, we can the receive
fast path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: #200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Whissi pushed a commit to Whissi/linux-stable that referenced this issue Jul 25, 2021
@gregkh
tcp: consistently disable header prediction for mptcp
ea66fcb 
commit 71158bb upstream.

The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when an MPTCP socket falls back to TCP, it can hit the TCP receive
fast-path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: multipath-tcp/mptcp_net-next#200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Whissi pushed a commit to Whissi/linux-stable that referenced this issue Jul 25, 2021
@gregkh
tcp: consistently disable header prediction for mptcp
1087278 
commit 71158bb upstream.

The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when an MPTCP socket falls back to TCP, it can hit the TCP receive
fast-path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: multipath-tcp/mptcp_net-next#200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
starnight pushed a commit to endlessm/linux that referenced this issue Sep 14, 2021
@smb49
tcp: consistently disable header prediction for mptcp
f858324 
BugLink: https://bugs.launchpad.net/bugs/1940139

commit 71158bb upstream.

The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when an MPTCP socket falls back to TCP, it can hit the TCP receive
fast-path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: multipath-tcp/mptcp_net-next#200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Stefan Bader <stefan.bader@canonical.com>
it-is-a-robot pushed a commit to openeuler-mirror/kernel that referenced this issue Oct 15, 2021
tcp: consistently disable header prediction for mptcp
14a0848 
stable inclusion
from stable-5.10.53
commit ea66fcb296058e0d41afa770f6900b335f8738a5
bugzilla: 175574 https://gitee.com/openeuler/kernel/issues/I4DTUX

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=ea66fcb296058e0d41afa770f6900b335f8738a5

--------------------------------

commit 71158bb upstream.

The MPTCP receive path is hooked only into the TCP slow-path.
The DSS presence allows plain MPTCP traffic to hit that
consistently.

Since commit e1ff9e8 ("net: mptcp: improve fallback to TCP"),
when an MPTCP socket falls back to TCP, it can hit the TCP receive
fast-path, and delay or stop triggering the event notification.

Address the issue explicitly disabling the header prediction
for MPTCP sockets.

Closes: multipath-tcp/mptcp_net-next#200
Fixes: e1ff9e8 ("net: mptcp: improve fallback to TCP")
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Chen Jun <chenjun102@huawei.com>
Acked-by: Weilong Chen <chenweilong@huawei.com>
Signed-off-by: Chen Jun <chenjun102@huawei.com>
Signed-off-by: Zheng Zengkai <zhengzengkai@huawei.com>
jenkins-tessares pushed a commit that referenced this issue Aug 11, 2023
selftests/bpf: fix a CI failure caused by vsock sockmap test
90f0074 
BPF CI has reported the following failure:

Error: #200/79 sockmap_listen/sockmap VSOCK test_vsock_redir
  Error: #200/79 sockmap_listen/sockmap VSOCK test_vsock_redir
  ./test_progs:vsock_unix_redir_connectible:1506: egress: write: Transport endpoint is not connected
  vsock_unix_redir_connectible:FAIL:1506
  ./test_progs:vsock_unix_redir_connectible:1506: ingress: write: Transport endpoint is not connected
  vsock_unix_redir_connectible:FAIL:1506
  ./test_progs:vsock_unix_redir_connectible:1506: egress: write: Transport endpoint is not connected
  vsock_unix_redir_connectible:FAIL:1506
  ./test_progs:vsock_unix_redir_connectible:1514: ingress: recv() err, errno=11
  vsock_unix_redir_connectible:FAIL:1514
  ./test_progs:vsock_unix_redir_connectible:1518: ingress: vsock socket map failed, a != b
  vsock_unix_redir_connectible:FAIL:1518
  ./test_progs:vsock_unix_redir_connectible:1525: ingress: want pass count 1, have 0

It’s because the recv(... MSG_DONTWAIT) syscall in the test case is
called before the queued work sk_psock_backlog() in the kernel finishes
executing. So the data to be read is still queued in psock->ingress_skb
and cannot be read by the user program. Therefore, the non-blocking
recv() reads nothing and reports an EAGAIN error.

So replace recv(... MSG_DONTWAIT) with xrecv_nonblock(), which calls
select() to wait for data to be readable or timeout before calls recv().

Fixes: d61bd8c ("selftests/bpf: add a test case for vsock sockmap")
Signed-off-by: Xu Kuohai <xukuohai@huawei.com>
Link: https://lore.kernel.org/r/20230804073740.194770-4-xukuohai@huaweicloud.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@pabeni pabeni

Labels
bug
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@pabeni @mjmartineau