Closed ilyacodes closed 6 years ago
Looks ok... I think if we transition to estab we should also kill any outstanding timers, but plink timer in ESTAB state is a no op so it doesn't actually matter. Would be nice to wrap delete+add in a helper function.
In many cases, we start a new timeout with
add_timeout
without checking whether another timeout is already running. For timers defined by 802.11s (represented by candidatet0
,t1
, andt2
timers), this is not the correct behaviour – they expect a single timer to be running at a time.Remove any existing timers with
rem_timeout
prior toadd_timeout
ing a new timer setting.Summary of behaviours:
authsae
, timer id is deleted when timer expireslibsae
MUST no-op removing an expired/non-existent timer0
—> timer id0
is reserved bylibsae
, removing it is no-op per 2.ii above