tls: missing reset of TCP connection established flag

[cspiel1]

The flag that indicates that the local TCP server to remote TCP client
connection was established has to be reset to false if the TCP connection is
closed.

Was found during working on: baresip/re#607

[cspiel1]

A valgrind test fails. Do we need Ubuntu 20.04 like in re baresip/re#606?
Why was 22.04 a problem in re?

Edit: I see. It can be reproduced when compiling re, rem and retest with clang. Then invoke:

valgrind --leak-check=full --show-reachable=yes --error-exitcode=42 ./retest -v -r
==6738== Memcheck, a memory error detector
==6738== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==6738== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info
==6738== Command: ./retest -v -r
==6738== 
### unhandled dwarf2 abbrev form code 0x25
### unhandled dwarf2 abbrev form code 0x25
### unhandled dwarf2 abbrev form code 0x25
### unhandled dwarf2 abbrev form code 0x1b
### unhandled dwarf2 abbrev form code 0x25
### unhandled dwarf2 abbrev form code 0x25
### unhandled dwarf2 abbrev form code 0x25
### unhandled dwarf2 abbrev form code 0x1b
==6738== Valgrind: debuginfo reader: ensure_valid failed:
==6738== Valgrind:   during call to ML_(img_get)
==6738== Valgrind:   request for range [18060034, +4) exceeds
==6738== Valgrind:   valid image size of 954048 for image:
==6738== Valgrind:   "/home/cspiel/src/retest/retest"
==6738== 
==6738== Valgrind: debuginfo reader: Possibly corrupted debuginfo file.
==6738== Valgrind: I can't recover.  Giving up.  Sorry.

[sreimers]

Yes we need to change the image or use -gdwarf-4 since valgrind has no dwarf-5 support.
And Ubuntu 22.04 fails with baresip/re#566

[cspiel1]

Do you want create a PR similar to baresip/re#606 for retest?