On 20/12/2023 10:25, Théo Lebrun wrote: > Hello, > > Thanks for your comments. I have a question for one: > > On Tue Dec 19, 2023 at 8:40 AM CET, Krzysztof Kozlowski wrote: >> On 18/12/2023 18:16, Théo Lebrun wrote: >>> Add DT-Schema bindings for the EyeQ5 reset controller. >>> >>> Signed-off-by: Théo Lebrun <theo.lebrun@xxxxxxxxxxx> >>> --- >>> .../bindings/reset/mobileye,eyeq5-reset.yaml | 69 +++++++++++++++++++ >>> MAINTAINERS | 2 + >>> include/dt-bindings/reset/mobileye,eyeq5-reset.h | 80 ++++++++++++++++++++++ >>> 3 files changed, 151 insertions(+) >>> > > [...] > >>> diff --git a/include/dt-bindings/reset/mobileye,eyeq5-reset.h b/include/dt-bindings/reset/mobileye,eyeq5-reset.h >>> new file mode 100644 >>> index 000000000000..ce59fe5409ac >>> --- /dev/null >>> +++ b/include/dt-bindings/reset/mobileye,eyeq5-reset.h >>> @@ -0,0 +1,80 @@ >>> +/* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */ >>> +/* >>> + * Copyright (C) 2023 Mobileye Vision Technologies Ltd. >>> + */ >>> + >>> +#ifndef _DT_BINDINGS_RESET_MOBILEYE_EYEQ5_RESET_H >>> +#define _DT_BINDINGS_RESET_MOBILEYE_EYEQ5_RESET_H >>> + >>> +/* Domain 0 */ >>> + >>> +/* 0..2 are reserved */ >> >> No, they are not. IDs cannot be reserved. IDs start from 0 and are >> incremented by 1. Reserving an ID contradicts to entire point of that >> ID, so either drop entire file or make this proper IDs. > > Those are hardware IDs. I get what you mean is that they should not leak There is no such thing as "hardware ID". It is some value passed to hardware/firmware or hardware register address/offset. There is no point to store hardware register values/offsets in the bindings. > into bindings. That implies a mapping operation from bindings IDs to > understood-by-hardware IDs. Can you confirm this is what you expect? https://en.wikipedia.org/wiki/Language_binding Bindings is an abstraction layer, not hardware, therefore my expectation is not having some sort of register values or offsets in the binding. Drop the header from bindings. Best regards, Krzysztof
