samples/net/zperf does not work for TCP in qemu_x86

[laperie]

Describe the bug
Running zperf in qemu gives IPv6 connect failed error

To Reproduce
Steps to reproduce the behavior:

1. mkdir build; cd build
2. cmake -DBOARD=qemu_x86
3. make
4. ensure iperf 2 is running as per instruction
5. Observe error

Expected behavior
No error

Impact
annoyance

Screenshots or console output

uart:~$ zperf tcp upload 2001:db8::2 5001 10 1K 1M
Remote port is 5001
Connecting to 2001:db8::2
Duration:	10.00 s
Packet size:	1024 bytes
Rate:		1024 kbps
Starting...
IPv6 connect failed (-60)

Environment (please complete the following information):

• Linux
• 0.10.0 SDK

Additional context
Running zperf with UDP works:

uart:~$ zperf udp upload 2001:db8::2 5001 10 1K 1M

Setting IP address 2001:db8::1
Setting destination IP address 2001:db8::2
Setting IP address 192.0.2.1
Setting destination IP address 192.0.2.2

Remote port is 5001
Connecting to 2001:db8::2
Duration:	10.00 s
Packet size:	1024 bytes
Rate:		1024 kbps
Starting...
Rate:		1.00 Mbps
nb_packets=128	delay=781168	adjust=81
nb_packets=256	delay=780972	adjust=-135
nb_packets=385	delay=781078	adjust=-93
nb_packets=514	delay=781147	adjust=-50
nb_packets=642	delay=781129	adjust=-45
nb_packets=770	delay=781108	adjust=8
nb_packets=899	delay=781183	adjust=38
nb_packets=1027	delay=780736	adjust=-413
nb_packets=1156	delay=781120	adjust=-16
-
Upload completed!
LAST PACKET NOT RECEIVED!!!
Statistics:		server	(client)
Duration:		0 us	(10.01 s)
Num packets:		0	(1280)
Num packets out order:	0
Num packets lost:	0
Jitter:			0 us
Rate:			0 Kbps	(1022 Kbps)

[jukkar]

The connect failed error would be fixed by #14648 or #14737 but after applying that one, the TCP engine is running out of memory and then crashing. I am investigating this new issue.

edit: added second alternative PR