On 13/09/2023 10:39, Moudy Ho (何宗原) wrote: >>> @@ -63,7 +64,6 @@ properties: >>> required: >>> - compatible >>> - reg >>> - - interrupts >> >> Why? commit msg tells nothing about it. Why interrupt is not erquired >> in >> mt8173? How dropping such requirement is anyhow related to mt8195? >> >> > The signals of the MDP engines are completely controlled by MTK's MUTEX > for starting and stopping frame processing, eliminating the need for > additional interrupts. Then it does not look like related to this patch at all. ... >>> >>> reg: >>> maxItems: 1 >>> @@ -50,17 +51,19 @@ properties: >>> maxItems: 1 >>> >>> clocks: >>> - items: >>> - - description: RDMA clock >>> - - description: RSZ clock >>> + oneOf: >>> + - items: >>> + - description: RDMA clock >>> + - description: SRAM shared component clock >>> + - items: >>> + - description: RDMA clock >> >> Why now mt8183 can have SRAM clock optional? How changing mt8183 is >> related to this patch? >> >> I'll finish the review, sorry fix basics here. >> >> Best regards, >> Krzysztof >> > The RDMA of only the 8183 needed to share SRMA with other component due > to the old desgin. > I attempted to describe both the situation of the 8183 and new designs > like the 8195, but it appears that this writing style may lead to > misunderstandings. > I am unsure if there are any ways to enhance it. Just like many other bindings, just look at Qualcomm, use allOf to restrict entries per variants. Best regards, Krzysztof