Hi Laurent,
On Mon, Dec 14, 2020 at 5:28 PM Laurent Pinchart laurent.pinchart@ideasonboard.com wrote:
On Mon, Dec 14, 2020 at 04:20:17PM +0000, Kieran Bingham wrote:
On 04/12/2020 22:01, Laurent Pinchart wrote:
devm_kcalloc() is the wrong API to allocate planes, as the lifetime of the planes is tied to the DRM device, not the device to driver binding. drmm_kcalloc() isn't a good option either, as it would result in the planes being freed before being unregistered during the managed cleanup of the DRM objects. Use a plain kcalloc(), and cleanup the planes and free the memory in the existing rcar_du_vsp_cleanup() handler.
Managed memory always seems to hurt - which is a shame, because it should be better throughout.
It's like we need a way to arbitrarily specify the lifetimes of objects correctly against another object... without being tied to a dev ...
I've been saying for years that devm_kzalloc() is a major regression. We've traded a memory leak for a use-after-free. The function has its use cases, there are objects that need to match the lifetime of the binding between a device and its driver, but that's a small minority.
https://en.wikipedia.org/wiki/The_law_of_conservation_of_misery
Gr{oetje,eeting}s,
Geert