[qca-nss-ecm] Add destroy re-try feature.
When a destroy request fails to sent to the acceleration
engine for some reason, that connection was being removed
from the ECM's database but the accel engine still keeps
that connection since it doesn't receive that request.
With this commit, if a destroy request fails, the defunct
timer is reset to a smaller timeout value and the defunct
callback is called again after the timeout expires.
Change-Id: I2e42a61234eaa53267aa0a987c417f0926f4d80c
Signed-off-by: Murat Sezgin <msezgin@codeaurora.org>
12 files changed