net/handshake: Fix handshake_req_destroy_test1

Recently, handshake_req_destroy_test1 started failing:

Expected handshake_req_destroy_test == req, but
    handshake_req_destroy_test == 0000000000000000
    req == 0000000060f99b40
not ok 11 req_destroy works

This is because "sock_release(sock)" was replaced with "fput(filp)"
to address a memory leak. Note that sock_release() is synchronous
but fput() usually delays the final close and clean-up.

The delay is not consequential in the other cases that were changed
but handshake_req_destroy_test1 is testing that handshake_req_cancel()
followed by closing the file actually does call the ->hp_destroy
method. Thus the PTR_EQ test at the end has to be sure that the
final close is complete before it checks the pointer.

We cannot use a completion here because if ->hp_destroy is never
called (ie, there is an API bug) then the test will hang.

Reported by: Guenter Roeck <[email protected]>
Closes: https://lore.kernel.org/netdev/[email protected]/T/#mac5c6299f86799f1c71776f3a07f9c566c7c3c40
Fixes: 4a0f07d ("net/handshake: Fix memory leak in __sock_create() and sock_alloc_file()")
Signed-off-by: Chuck Lever <[email protected]>
Reviewed-by: Hannes Reinecke <[email protected]>
Link: https://lore.kernel.org/r/170724699027.91401.7839730697326806733.stgit@oracle-102.nfsv4bat.org
Signed-off-by: Jakub Kicinski <[email protected]>

net/handshake/handshake-test.c

@@ -471,7 +471,10 @@ static void handshake_req_destroy_test1(struct kunit *test)
 	handshake_req_cancel(sock->sk);
 
 	/* Act */
-	fput(filp);
+	/* Ensure the close/release/put process has run to
+	 * completion before checking the result.
+	 */
+	__fput_sync(filp);
 
 	/* Assert */
 	KUNIT_EXPECT_PTR_EQ(test, handshake_req_destroy_test, req);