On Wed, Dec 18, 2019 at 06:57:10PM +0000, Saleem, Shiraz wrote: > > > > Subject: Re: [PATCH v3 04/20] i40e: Register a virtbus device to > > > > provide RDMA > [.....] > > > > > > > > > And who owns the memory of this thing that is supposed to be > > > > dynamically controlled by something OUTSIDE of this driver? Who > > > > created that thing 3 pointers deep? What happens when you leak the > > > > memory below (hint, you did), and who is supposed to clean it up if > > > > you need to properly clean it up if something bad happens? > > > > > > The i40e_info object memory is tied to the PF driver. > > > > What is a "PF"? > > physical function. > > > > > > The object hierarchy is, > > > > > > i40e_pf: pointer to i40e_client_instance > > > ----- i40e_client_instance: i40e_info > > > ----- i40e_info: virtbus_device > > > > So you are 3 pointers deep to get a structure that is dynamically controlled? Why > > are those "3 pointers" not also represented in sysfs? > > You have a heiarchy within the kernel that is not being represented that way to > > userspace, why? > > > > Hint, I think this is totally wrong, you need to rework this to be sane. > > > > > For each PF, there is a client_instance object allocated. > > > > Great, make it dynamic and in the device tree. > > > > > The i40e_info object is populated and the virtbus_device hanging off this object > > is registered. > > > > Great, make that dynamic and inthe device tree. > > > > If you think this is too much, then your whole mess here is too much and needs to > > be made a lot simpler. > > > > I think we can decouple the virtbus_device object from i40e_info object. > > Instead allocate a i40e_virtbus_device object which contains > the virtbus_device and a pointer to i40e_info object for the > RDMA driver to consume on probe(). Register it the virtbus, and provide > a release callback to free up its memory. > > Sending a patch snippet to hopefully make it clearer. Yes, this looks a little bit more sane. greg k-h
