Re: [PATCH net v2 1/2] tcp: process the 3rd ACK with sk_socket for TFO/MPTCP

[Matthieu Baerts <matttbe@kernel.org>]

Hi Eric,

On 23/07/2024 17:38, Eric Dumazet wrote:
> On Tue, Jul 23, 2024 at 4:58 PM Matthieu Baerts <matttbe@kernel.org> wrote:
>>
>> Hi Eric,
>>
>> +cc Neal
>> -cc Jerry (NoSuchUser)
>>
>> On 23/07/2024 16:37, Eric Dumazet wrote:
>>> On Thu, Jul 18, 2024 at 12:34 PM Matthieu Baerts (NGI0)
>>> <matttbe@kernel.org> wrote:
>>>>
>>>> The 'Fixes' commit recently changed the behaviour of TCP by skipping the
>>>> processing of the 3rd ACK when a sk->sk_socket is set. The goal was to
>>>> skip tcp_ack_snd_check() in tcp_rcv_state_process() not to send an
>>>> unnecessary ACK in case of simultaneous connect(). Unfortunately, that
>>>> had an impact on TFO and MPTCP.
>>>>
>>>> I started to look at the impact on MPTCP, because the MPTCP CI found
>>>> some issues with the MPTCP Packetdrill tests [1]. Then Paolo suggested
>>>> me to look at the impact on TFO with "plain" TCP.
>>>>
>>>> For MPTCP, when receiving the 3rd ACK of a request adding a new path
>>>> (MP_JOIN), sk->sk_socket will be set, and point to the MPTCP sock that
>>>> has been created when the MPTCP connection got established before with
>>>> the first path. The newly added 'goto' will then skip the processing of
>>>> the segment text (step 7) and not go through tcp_data_queue() where the
>>>> MPTCP options are validated, and some actions are triggered, e.g.
>>>> sending the MPJ 4th ACK [2] as demonstrated by the new errors when
>>>> running a packetdrill test [3] establishing a second subflow.
>>>>
>>>> This doesn't fully break MPTCP, mainly the 4th MPJ ACK that will be
>>>> delayed. Still, we don't want to have this behaviour as it delays the
>>>> switch to the fully established mode, and invalid MPTCP options in this
>>>> 3rd ACK will not be caught any more. This modification also affects the
>>>> MPTCP + TFO feature as well, and being the reason why the selftests
>>>> started to be unstable the last few days [4].
>>>>
>>>> For TFO, the existing 'basic-cookie-not-reqd' test [5] was no longer
>>>> passing: if the 3rd ACK contains data, and the connection is accept()ed
>>>> before receiving them, these data would no longer be processed, and thus
>>>> not ACKed.
>>>>
>>>> One last thing about MPTCP, in case of simultaneous connect(), a
>>>> fallback to TCP will be done, which seems fine:
>>>>
>>>>   `../common/defaults.sh`
>>>>
>>>>    0 socket(..., SOCK_STREAM|SOCK_NONBLOCK, IPPROTO_MPTCP) = 3
>>>>   +0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress)
>>>>
>>>>   +0 > S  0:0(0)                 <mss 1460, sackOK, TS val 100 ecr 0,   nop, wscale 8, mpcapable v1 flags[flag_h] nokey>
>>>>   +0 < S  0:0(0) win 1000        <mss 1460, sackOK, TS val 407 ecr 0,   nop, wscale 8, mpcapable v1 flags[flag_h] nokey>
>>>>   +0 > S. 0:0(0) ack 1           <mss 1460, sackOK, TS val 330 ecr 0,   nop, wscale 8, mpcapable v1 flags[flag_h] nokey>
>>>>   +0 < S. 0:0(0) ack 1 win 65535 <mss 1460, sackOK, TS val 700 ecr 100, nop, wscale 8, mpcapable v1 flags[flag_h] key[skey=2]>
>>>>
>>>>   +0 write(3, ..., 100) = 100
>>>>   +0 >  . 1:1(0)     ack 1 <nop, nop, TS val 845707014 ecr 700, nop, nop, sack 0:1>
>>>>   +0 > P. 1:101(100) ack 1 <nop, nop, TS val 845958933 ecr 700>
>>>>
>>>> Simultaneous SYN-data crossing is also not supported by TFO, see [6].
>>>>
>>>> Link: https://github.com/multipath-tcp/mptcp_net-next/actions/runs/9936227696 [1]
>>>> Link: https://datatracker.ietf.org/doc/html/rfc8684#fig_tokens [2]
>>>> Link: https://github.com/multipath-tcp/packetdrill/blob/mptcp-net-next/gtests/net/mptcp/syscalls/accept.pkt#L28 [3]
>>>> Link: https://netdev.bots.linux.dev/contest.html?executor=vmksft-mptcp-dbg&test=mptcp-connect-sh [4]
>>>> Link: https://github.com/google/packetdrill/blob/master/gtests/net/tcp/fastopen/server/basic-cookie-not-reqd.pkt#L21 [5]
>>>> Link: https://github.com/google/packetdrill/blob/master/gtests/net/tcp/fastopen/client/simultaneous-fast-open.pkt [6]
>>>> Fixes: 23e89e8ee7be ("tcp: Don't drop SYN+ACK for simultaneous connect().")
>>>> Suggested-by: Paolo Abeni <pabeni@redhat.com>
>>>> Suggested-by: Kuniyuki Iwashima <kuniyu@amazon.com>
>>>> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
>>>> ---
>>>> Notes:
>>>>  - We could also drop this 'goto consume', and send the unnecessary ACK
>>>>    in this simultaneous connect case, which doesn't seem to be a "real"
>>>>    case, more something for fuzzers. But that's not what the RFC 9293
>>>>    recommends to do.
>>>>  - v2:
>>>>    - Check if the SYN bit is set instead of looking for TFO and MPTCP
>>>>      specific attributes, as suggested by Kuniyuki.
>>>>    - Updated the comment above
>>>>    - Please note that the v2 has been sent mainly to satisfy the CI (to
>>>>      be able to catch new bugs with MPTCP), and because the suggestion
>>>>      from Kuniyuki looks better. It has not been sent to urge TCP
>>>>      maintainers to review it quicker than it should, please take your
>>>>      time and enjoy netdev.conf :)
>>>> ---
>>>>  net/ipv4/tcp_input.c | 7 ++++++-
>>>>  1 file changed, 6 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
>>>> index ff9ab3d01ced..bfe1bc69dc3e 100644
>>>> --- a/net/ipv4/tcp_input.c
>>>> +++ b/net/ipv4/tcp_input.c
>>>> @@ -6820,7 +6820,12 @@ tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb)
>>>>                 if (sk->sk_shutdown & SEND_SHUTDOWN)
>>>>                         tcp_shutdown(sk, SEND_SHUTDOWN);
>>>>
>>>> -               if (sk->sk_socket)
>>>> +               /* For crossed SYN cases, not to send an unnecessary ACK.
>>>> +                * Note that sk->sk_socket can be assigned in other cases, e.g.
>>>> +                * with TFO (if accept()'ed before the 3rd ACK) and MPTCP (MPJ:
>>>> +                * sk_socket is the parent MPTCP sock).
>>>> +                */
>>>> +               if (sk->sk_socket && th->syn)
>>>>                         goto consume;
>>>
>>> I think we should simply remove this part completely, because we
>>> should send an ack anyway.
>>
>> Thank you for having looked, and ran the full packetdrill test suite!
>>
>>>
>>> diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
>>> index ff9ab3d01ced89570903d3a9f649a637c5e07a90..91357d4713182078debd746a224046cba80ea3ce
>>> 100644
>>> --- a/net/ipv4/tcp_input.c
>>> +++ b/net/ipv4/tcp_input.c
>>> @@ -6820,8 +6820,6 @@ tcp_rcv_state_process(struct sock *sk, struct
>>> sk_buff *skb)
>>>                 if (sk->sk_shutdown & SEND_SHUTDOWN)
>>>                         tcp_shutdown(sk, SEND_SHUTDOWN);
>>>
>>> -               if (sk->sk_socket)
>>> -                       goto consume;
>>>                 break;
>>>
>>>         case TCP_FIN_WAIT1: {
>>>
>>>
>>> I have a failing packetdrill test after  Kuniyuki  patch :
>>>
>>>
>>>
>>> //
>>> // Test the simultaneous open scenario that both end sends
>>> // SYN/data. Although we don't support that the connection should
>>> // still be established.
>>> //
>>> `../../common/defaults.sh
>>>  ../../common/set_sysctls.py /proc/sys/net/ipv4/tcp_timestamps=0`
>>>
>>> // Cache warmup: send a Fast Open cookie request
>>>     0 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
>>>    +0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0
>>>    +0 sendto(3, ..., 0, MSG_FASTOPEN, ..., ...) = -1 EINPROGRESS
>>> (Operation is now in progress)
>>>    +0 > S 0:0(0) <mss 1460,nop,nop,sackOK,nop,wscale 8,FO,nop,nop>
>>>  +.01 < S. 123:123(0) ack 1 win 14600 <mss
>>> 1460,nop,nop,sackOK,nop,wscale 6,FO abcd1234,nop,nop>
>>>    +0 > . 1:1(0) ack 1
>>>  +.01 close(3) = 0
>>>    +0 > F. 1:1(0) ack 1
>>>  +.01 < F. 1:1(0) ack 2 win 92
>>>    +0 > .  2:2(0) ack 2
>>>
>>>
>>> //
>>> // Test: simulatenous fast open
>>> //
>>>  +.01 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
>>>    +0 fcntl(4, F_SETFL, O_RDWR|O_NONBLOCK) = 0
>>>    +0 sendto(4, ..., 1000, MSG_FASTOPEN, ..., ...) = 1000
>>>    +0 > S 0:1000(1000) <mss 1460,nop,nop,sackOK,nop,wscale 8,FO
>>> abcd1234,nop,nop>
>>> // Simul. SYN-data crossing: we don't support that yet so ack only remote ISN
>>> +.005 < S 1234:1734(500) win 14600 <mss 1040,nop,nop,sackOK,nop,wscale
>>> 6,FO 87654321,nop,nop>
>>>    +0 > S. 0:0(0) ack 1235 <mss 1460,nop,nop,sackOK,nop,wscale 8>
>>>
>>> // SYN data is never retried.
>>> +.045 < S. 1234:1234(0) ack 1001 win 14600 <mss
>>> 940,nop,nop,sackOK,nop,wscale 6,FO 12345678,nop,nop>
>>>    +0 > . 1001:1001(0) ack 1
>>
>> I recently sent a PR -- already applied -- to Neal to remove this line:
>>
>>   https://github.com/google/packetdrill/pull/86
>>
>> I thought it was the intension of Kuniyuki's patch not to send this ACK
>> in this case to follow the RFC 9293's recommendation. This TFO test
>> looks a bit similar to the example from Kuniyuki's patch:
>>
>>
>> --------------- 8< ---------------
>>  0 socket(..., SOCK_STREAM|SOCK_NONBLOCK, IPPROTO_TCP) = 3
>> +0 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress)
>>
>> +0 > S  0:0(0) <mss 1460,sackOK,TS val 1000 ecr 0,nop,wscale 8>
>> +0 < S  0:0(0) win 1000 <mss 1000>
>> +0 > S. 0:0(0) ack 1 <mss 1460,sackOK,TS val 3308134035 ecr 0,nop,wscale 8>
>> +0 < S. 0:0(0) ack 1 win 1000
>>
>>   /* No ACK here */
>>
>> +0 write(3, ..., 100) = 100
>> +0 > P. 1:101(100) ack 1
>> --------------- 8< ---------------
>>
>>
>>
>> But maybe here that should be different for TFO?
>>
>> For my case with MPTCP (and TFO), it is fine to drop this 'goto consume'
>> but I don't know how "strict" we want to be regarding the RFC and this
>> marginal case.
> 
> Problem of this 'goto consume' is that we are not properly sending a
> DUPACK in this case.
> 
>  +.01 socket(..., SOCK_STREAM, IPPROTO_TCP) = 4
>    +0 fcntl(4, F_SETFL, O_RDWR|O_NONBLOCK) = 0
>    +0 sendto(4, ..., 1000, MSG_FASTOPEN, ..., ...) = 1000
>    +0 > S 0:1000(1000) <mss 1460,nop,nop,sackOK,nop,wscale 8,FO
> abcd1234,nop,nop>
> // Simul. SYN-data crossing: we don't support that yet so ack only remote ISN
> +.005 < S 1234:1734(500) win 14600 <mss 1040,nop,nop,sackOK,nop,wscale
> 6,FO 87654321,nop,nop>
>    +0 > S. 0:0(0) ack 1235 <mss 1460,nop,nop,sackOK,nop,wscale 8>
> 
> +.045 < S. 1234:1234(0) ack 1001 win 14600 <mss
> 940,nop,nop,sackOK,nop,wscale 6,FO 12345678,nop,nop>
>    +0 > . 1001:1001(0) ack 1 <nop,nop,sack 0:1>  // See here

I'm sorry, but is it normal to have 'ack 1' with 'sack 0:1' here?

> Not sending a dupack seems wrong and could hurt.

Indeed, I thought the RFC 9293 was not allowing that, but I didn't see
anything forbidding this DUPACK:

  https://www.rfc-editor.org/rfc/rfc9293.html#section-3.5-7

> I had reservations about this part, if you look back to the discussion.
> 
> This is why Kuniyuki added in his changelog :
> 
>     Note that tcp_ack_snd_check() in tcp_rcv_state_process() is skipped not to
>     send an unnecessary ACK, but this could be a bit risky for net.git, so this
>     targets for net-next.

I understand, I can send a v3 dropping this part, and not including
patch 2/2 for -net. I can also send a PR to Neal re-adding the ACK with
'sack' (if it is normal).

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.