Re: [tcpm] 793bis ready to go?

[Michael Tuexen <[e-mail]>]

> On 21. Feb 2021, at 15:33, Scheffenegger, Richard <[e-mail]> wrote:
> 
> 
> Well, making the transmission of the ACK-only-the-SYN optional when data
> is present, which needs to be ACK'd anyway..
> 
> That way, current FBSD and Linux behaviour both would comply with
> 793bis, while I agree with Michael T. that the behavior of FBSD is more
> efficient.
Hi Richard,

I'm not sure about your statement about compliance of the Linux stack.
My testing show that the following packetdrill scripts succeeds:

 0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+0.000 fcntl(3, F_GETFL) = 0x2 (flags O_RDWR)
+0.000 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+0.000 connect(3, ..., ...) = -1 EINPROGRESS (Operation now in progress)
+0.000 > S    0:0(0)             win 64240 <mss 1460,sackOK,TS val 100 ecr 0,nop,wscale 7>
+0.050 < S.   0:512(512) ack   1 win 65535 <mss 1460,sackOK,eol,eol>
+0.000 >  .   1:1(0)     ack   1 win 64240
+0.000 recv(3, ..., 1024, 0) = -1 EWOULDBLOCK (Operation would block)
+0.000 close(3) = 0
+0.000 > F.   1:1(0)     ack   1 win 64240
+0.000 < F.   1:1(0)     ack   2 win 65535
+0.000 >  .   2:2(0)     ack   2 win 64239

This means that the data on the SYN-ACK is simply dropped. I don't think
this is the handling described in RFC 793.

MacOS handles the case similar to FreeBSD.

Best regards
Michael 
> 
> Richard
> 
> Am 20.02.2021 um 22:53 schrieb Michael Tuexen:
>> 
>> 
>>> On 20. Feb 2021, at 22:37, Joseph Touch <[e-mail]> wrote:
>>> 
>>> Even if you prefer the FreeBSD variant, the text below doesn’t appear to fix this; it will end up sending the SYN-ACK of the received data *and a second ACK* at the end of the sixth step.
>> I think the idea of Richard is that if the text is
>> 
>>          If SND.UNA > ISS (our SYN has been ACKed), change the
>>          connection state to ESTABLISHED, form an ACK segment
>>          <SEQ=SND.NXT><ACK=RCV.NXT><CTL=ACK>
>>          and should send it.
>> 
>> Then NOT following the should, but only send the ACK at the end of step 6.
>> Wouldn't that work?
>> 
>> Best regards
>> Michael
>>> 
>>> Joe
>>> 
>>>> On Feb 20, 2021, at 12:56 PM, [e-mail] wrote:
>>>> 
>>>>> On 18. Feb 2021, at 14:59, Scheffenegger, Richard <[e-mail]> wrote:
>>>>> 
>>>>> One more nit, around dealing with data-in-SYN,ACK
>>>>> 
>>>>> 793bis (page 71) has this text, which unfortunately doesn have any
>>>>> RFC2119 language.
>>>>> 
>>>>> But in short, it seems to require the transmission of a pure ACK just
>>>>> for the SYN bit, and a 2nd ACK right afterwards (step 6) to acknowledge
>>>>> the data.
>>>>> 
>>>>> The FBSD implementation was found to skip over the ACK-only-the-SYN
>>>>> intermediate step (which must be acceptable, as that ACK may get lost in
>>>>> the network), sending out only the cummulative ACK for SYN and data,
>>>>> after having transitioned into ESTABLISHED.
>>>> Hi,
>>>> 
>>>> I really prefer the FreeBSD behaviour. A TCP stack should not need to send
>>>> two segments at the same time when handling the event of a single segment reception.
>>>> 
>>>> Best regards
>>>> Michael, as an individual
>>>>> 
>>>>> But strictly speaking, this behavior is a deviation from 793 / 793bis.
>>>>> 
>>>>> Perhaps a "and should send it" would do, to allow this behavior,
>>>>> conserving one pure ACK...
>>>>> 
>>>>> Best regards,
>>>>> Richard
>>>>> 
>>>>> 
>>>>> 
>>>>> If the state is SYN-SENT then
>>>>> 
>>>>>       first check the ACK bit
>>>>> 
>>>>>          If the ACK bit is set
>>>>> :
>>>>> :
>>>>> fourth check the SYN bit
>>>>> 
>>>>>          This step should be reached only if the ACK is ok, or there
>>>>>          is no ACK, and it the segment did not contain a RST.
>>>>> 
>>>>>          If the SYN bit is on and the security/compartment is
>>>>>>>>       acceptable then, RCV.NXT is set to SEG.SEQ+1, IRS is set to
>>>>>          SEG.SEQ.  SND.UNA should be advanced to equal SEG.ACK (if
>>>>>          there is an ACK), and any segments on the retransmission
>>>>>          queue that are thereby acknowledged should be removed.
>>>>> 
>>>>>          If SND.UNA > ISS (our SYN has been ACKed), change the
>>>>>          connection state to ESTABLISHED, form an ACK segment
>>>>> 
>>>>>>>>          <SEQ=SND.NXT><ACK=RCV.NXT><CTL=ACK>
>>>>> 
>>>>>          and send it.  Data or controls that were queued for
>>>>>          transmission may be included.  If there are other controls
>>>>> or text in the segment then continue processing at the sixth
>>>>>          step below where the URG bit is checked, otherwise return.
>>>>> 
>>>>> _______________________________________________
>>>>> tcpm mailing list
>>>>> [e-mail]
>>>>> https://www.ietf.org/mailman/listinfo/tcpm
>>>> 
>>>> _______________________________________________
>>>> tcpm mailing list
>>>> [e-mail]
>>>> https://www.ietf.org/mailman/listinfo/tcpm
>>> 
>>> _______________________________________________
>>> tcpm mailing list
>>> [e-mail]
>>> https://www.ietf.org/mailman/listinfo/tcpm
>> 
>> _______________________________________________
>> tcpm mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/tcpm
>>