Thank you for providing a patch! That patch looks straight forward and (almost) resembles the upstream change, obviously missing the parts that do not yet exist in systemd v245.
The only other thing I found missing is an upstream change in src/resolve/resolved-dns-transaction.c:
I've created a new patch for Focal, including this change and also adopted the patch to Hirsute.
I have a few open questions:
1/ Did you leave out that resolved-dns-transaction.c change on purpose?
2/ What is the best way to reproduce this issue? Can it somehow be triggered?
3/ How to test/confirm that the issue is indeed fixed? Is there any way other than observing for 24 hours?
Thank you for providing a patch! That patch looks straight forward and (almost) resembles the upstream change, obviously missing the parts that do not yet exist in systemd v245.
The only other thing I found missing is an upstream change in src/resolve/ resolved- dns-transaction .c:
``` _stop_timeout( DnsTransaction *t) {
static void dns_transaction
assert(t);
- t->timeout_ event_source = sd_event_ source_ unref(t- >timeout_ event_source) ; event_source = sd_event_ source_ disable_ unref(t- >timeout_ event_source) ;
+ t->timeout_
}
DnsTransaction* dns_transaction _free(DnsTransa ction *t) {
```
I've created a new patch for Focal, including this change and also adopted the patch to Hirsute.
I have a few open questions: dns-transaction .c change on purpose?
1/ Did you leave out that resolved-
2/ What is the best way to reproduce this issue? Can it somehow be triggered?
3/ How to test/confirm that the issue is indeed fixed? Is there any way other than observing for 24 hours?