[qca-nss-ecm] Fix destroy failure handler issue

There is race between two defunct calls, if the first
one is failed and while we are handling it, another
defunct call comes before setting the retry timer group.
This patch avoids to call another defunct callback
while in the middle of handling the failed one.

Signed-off-by: Murat Sezgin <msezgin@codeaurora.org>
Change-Id: I7f6799499b3d6ee9b86df349136eb9187a5fb7eb
2 files changed