Hello,
>
> pf_test_rule uses the skw and sks pointers after pf_state_insert sets
> them via pf_create_state. i would happily change pf_test rule so it
> reads the pf_state_key pointers out of pf_state rather than carry them
> around on the stack like that, but i figured the diff was big enough as
> it was.
I've forgot about pf_test_rule(). perhaps just setting skw/sks
to NULL on failure (with reference gone) should be sufficient
at this point.
>
> > the recent diff is OK by me.
>
> thanks. i'll wait until i can be around to look after it before i commit
> it.
thanks and
regards
sashan
