What were you trying to do that didn't work?

Request to backport upstream fix 
https://bugs.passt.top/show_bug.cgi?id=197

As per the existing code https://passt.top/passt/tree/udp_flow.c#n20 included in RHEL, udp connect timeout was hardcoded as 180 seconds

#define UDP_CONN_TIMEOUT    180 /* s, timeout for ephemeral or local bind */

When we have more containers running on a node and if an administrator runs a netstat or ss output, they may end up seeing 1000's of connection in this state for 180 seconds. Technically this is not a cause of concern and actually harmless, and it's the expected behaviour, but it looks exceedingly noisy in the output of ss(8) and might worry admins as a result

udp        0      0 192.168.40.183:33750      192.18.70.67:53          ESTABLISHED 2219       300026401  3545482/pasta
udp        0      0 192.168.40.183:33835      192.18.70.33:53          ESTABLISHED 2219       300034434  3545482/pasta

root@node:~# netstat -laputen | grep pasta | wc
    845    7605  103935
root@node:~# netstat -laputen | grep pasta | wc
    854    7686  105042
    
These flows have a default hardcoded expiry of 180 seconds    
    
What is the impact of this issue to you?

No impact but its noisy

Please provide the package NVR for which the bug is seen:
 passt-0^20250512.g8ec1341-4.el9_7.

How reproducible is this bug?:
Always

Steps to reproduce

1. Run a rootless container
2. Inside a container run, curl -Lv http://redhat.com a 100 times
3. On the host, check ss/netstat output

     netstat -anp |grep pasta

Expected results

Reduce the default timeout of 180 seconds. Make it configurable or adjust timeouts based on datagrams /with without reply and with a max sealing(less than 180 seconds)

Actual results

Default UDP timeout is hardcoded as 180 seconds