The point of my comment was that the connection between CMEs and flares would make for a good 20 page research paper in a college course. The connections are still only understood at a kind of cartoon level, after decades of concerted research. But I can tell you what the basic cartoon is, and you can think of that as a kind of operating hypothesis that helps organize many of the observations, but is still a long way from a predictive model. The basic idea is that the energy for flares and for CMEs both seem to come from magnetic free energy stored in the corona due to the magnetic and turbulent nature of the convection zone. The magnetic field is capable of trapping pockets of hot gas that has been energized by this magnetic activity, but then when the field lines get significantly twisted up they reach a point where they can release that energy and relax to a smoother configuration. This relaxation does two things-- it releases energy, and it reconfigures the field such that it can release these pockets of trapped plasma. The former is the solar flare, the latter is the CME, and that's why the two often come together.
However, there seems to be a full range of weak to strong flares, and weak ones would not be associated with CMEs. On top of all that, there may be other processes that heat the corona, and other processes that lead to the solar wind, so one big question is if the heating of the corona is mostly from lots and lots of small flares, and the solar wind is lots and lots of small CMEs, or if the general heating of the corona and general driving of the wind are separate processes from flares and CMEs.
Also, you mention closed and open field regions, so these questions can be asked separately in those two domains. But CMEs would normally come from closed field regions that open up when the field reconfiguration occurs, I believe open field regions tend to be rather sedate and not associated with either strong flares or CMEs, but rather a more steady wind. In the closed field regions, CMEs are often associated with features called "helmet streamers," which are regions of trapped plasma that are kind of combed out at their highest point by the solar wind, creating an interface between closed and open fields that is a good setup for getting field reconfiguration that could launch out some plasma.