As I understand, each clock definition corresponds to the clock CG settings provided to different hardware, and each hardware driver can control its own clock CG through the CCF to control their CG in clock driver. So they can be an abstraction values between driver and DTS. Similarly, the GCE subsys ID and GCE event ID correspond to symbols used by GCE to control various hardware, and each hardware driver can use these IDs to generate commands buffer for GCE through the API provided by the GCE driver and achieve the desired control over their hardware. I guess the difference is that the clock driver has a platform-specific clock table to store these binding headers, while the GCE driver does not have a platform-specific thread priority table, subsys ID table, and event ID table. Instead, the GCE client drivers can directly obtain their respective hardware settings from the DTS. On the other hand, definitions like CLK_TOP_MAINPLL_D3, CLK_TOP_MAINPLL_D4, etc., correspond to different clock frequency divider levels, and the CMDQ_THR_PRIO_X for GCE thread priority also corresponds to different priority levels for GCE threads. Therefore, I am not quite sure why GCE thread priority cannot be considered a binding when it is also a symbol number for a hardware level setting. If the condition for becoming a binding header is that it `must` be used by a specific driver, such as a platform-specific table, then I will remove the entire GCE dt-binding header. Because the current usage of these definitions is that each GCE client drivers can directly store these GCE definitions through the DTS, just like IRQ IDs, and without the need for an additional table defined by the GCE driver. I am not unwilling to change the code, but I hope to first understand what qualifies as a binding header. This way, I can confidently remove the MT8196 GCE binding header, and other developers will also know that the GCE binding header for the previous MTXXXX is not needed. I sincerely hope you can guide me to meet your expectations and standards. Regards, Jason-JH.Lin > > > Â Â Â Â ``` > > Best regards, > Krzysztof >
