TCP question: Is this simultaneous close handling broken?

[Peter Wemm]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 1/7/14, 7:07 PM, Peter Wemm wrote:
> On 1/7/14, 3:39 PM, Mike Tancsa wrote:
>> On 1/7/2014 5:40 PM, Peter Wemm wrote:
>> 
>>> The packet may be dropped without processing the FIN flag.
>> 
>>> MFC after:	never
>> 
>> Hi, Are there any potential side effects to this fix ? The original 
>> author said they were not going to MFC due to possible regressions.
>> I know you probably see more FreeBSD traffic then most at Y!, and so
>> are very sensitive to this, but thought I would ask for
>> clarification.
>> 
>> ---Mike
> 
> Actually, I'm very troubled by that entire chunk of code.

I think the correct fix is to back out r239672 from rrs, and modify
r258821 from the PR so that it understands that that it applies to only
the first FIN packet we get.  I slightly moved the test for clarity and
for room to comment.

http://people.freebsd.org/~peter/tcp_input.c.diff

I believe that eadler's r258821 interferes with normal cc operation for a
small window after the remote has sent a FIN. Rev r258821 also turns
r239672 into dead code.
- -- 
Peter Wemm - peter at wemm.org; peter at FreeBSD.org; peter at yahoo-inc.com; KI6FJV
UTF-8: for when a ' just won\342\200\231t do.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (Darwin)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iEYEARECAAYFAlLNpI4ACgkQFRKuUnJ3cX9yHwCfUVVXcsfHtKRfsCeQ1OVksAYW
FskAn3PuJozJw0kVKpfJuaEoHBOClTdY
=63Li
-----END PGP SIGNATURE-----