systemctl: fix continue placement in clean-or-freeze error handling

systemctl: fix continue placement in clean-or-freeze error handling

When sd_bus_call() fails, the continue was inside the
'if (ret == EXIT_SUCCESS)' guard, so only the first failure skipped
adding the unit to the job waiter. On the second and subsequent
failures, the unit was still passed to bus_wait_for_units_add_unit()
despite no job being started, causing bus_wait_for_units_run() to
hang indefinitely.

Move continue outside the guard so any failure skips the waiter
registration. The guard still prevents ret from being overwritten by
a later error code.

Signed-off-by: dongshengyuan <dongshengyuan@uniontech.com>