Hello Giannis,
On Thu, Jul 06, 2023 at 01:54:36PM +0300, Kapetanakis Giannis wrote:
</snip>
>
> Hi sashan,
>
> Your logic seems fine to me. Indeed it acts like they are persistent. They
> don't have the p flag.
>
> # pfctl -a 'relayd/dir-imap' -sT -vg
> --a-r-- dir-imap?????????????? relayd/dir-imap
> # pfctl -a 'relayd/dir-sieve' -sT -vg
> --a-r-- dir-sieve???????????? relayd/dir-sieve
> # pfctl -a 'relayd/dir-lmtp' -sT -vg
> --a-r-- dir-lmtp?????????????? relayd/dir-lmtp
>
> note that in pfe_filter.c in init_tables() is see
>
> 73: tables[i].pfrt_flags |= PFR_TFLAG_PERSIST;
>
> Giannis
> ps. I've send a diff about statistics in tech@
> It does more than just checking table->up status, but your approach is much
> better.
>
thanks for checking. so this needs fix. the persistent
flag (PFR_TFLAG_PERSIST) gets lost if ioctl() loads rules
before table. pfctl(8) also sets PFR_TFLAG_PERSIST on all tables
it creates/loads. But the flag gets lost if table is created
by rule beforehand. See my earlier email.
this will need some more though. I'm not sure if it is a bug
in pf(4) itself or feature of pf(4). If it is pf(4) feature,
then fix needs to go to relayd.
thank you for all your help here.
regards
sashan
