tipc: avoid packets leaking on socket receive queue
Even if we drain receive queue thoroughly in tipc_release() after tipc socket is removed from rhashtable, it is possible that some packets are in flight because some CPU runs receiver and did rhashtable lookup before we removed socket. They will achieve receive queue, but nobody delete them at all. To avoid this leak, we register a private socket destructor to purge receive queue, meaning releasing packets pending on receive queue will be delayed until the last reference of tipc socket will be released. Signed-off-by:Ying Xue <ying.xue@windriver.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
parent
3c25a860
Please register or sign in to comment