Yogender Arya
2023-Feb-01 07:04 UTC
[fdo] Necessity for using the component helper library for DRM drivers.
Hi, I am new to the Linux and DRM framework in Linux. I was going through some reference drivers like VC4, exynos, etnaviv, Renesas (rcar-du), rockchip. Almost in all these cases, they use a component helpers to glue together the different IPs under the hood of a pseudo master driver. It should also be possible to let the individual components probe on their own, and they individually register with the DRM core to provide the required functionality. If an SoC contains different IP blocks to render and driver the display, should we not go with the basic probing for each individual IP and then register with the DRM core, rather than bringing it under one hood. Are there any inherent benefits for doing things the other way? Sorry for asking if it is too obvious. Thanks Yogender Arya