Re: [PATCH v6 04/12] dt-bindings: reset: nuvoton: add binding for ma35d1 IP reset control

[Jacky Huang <ychuang570808@gmail.com>]

Dear Krzysztof,



On 2023/3/29 下午 05:27, Krzysztof Kozlowski wrote:
> On 29/03/2023 11:12, Jacky Huang wrote:
>>>> +        };
>>>> +    };
>>>> +...
>>>> +
>>>> diff --git a/include/dt-bindings/reset/nuvoton,ma35d1-reset.h b/include/dt-bindings/reset/nuvoton,ma35d1-reset.h
>>>> new file mode 100644
>>>> index 000000000000..f3088bc0afec
>>>> --- /dev/null
>>>> +++ b/include/dt-bindings/reset/nuvoton,ma35d1-reset.h
>>>> @@ -0,0 +1,108 @@
>>>> +/* SPDX-License-Identifier: (GPL-2.0+ OR BSD-2-Clause) */
>>> Weird license, why 2.0+?
>>>
>>> You already got here comment about license in previous version of this
>>> patch.
>>>
>> In fact, I always be confused with the license.
>> It should be right if I just copy one from the same directory
>> /include/dt-bindings/reset/.
>> But, when I find there are several versions there, I was confused and
>> just select one of them.
>>
>> /* SPDX-License-Identifier: GPL-2.0-only */
>> /* SPDX-License-Identifier: GPL-2.0+ */
>> /* SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause)*/
>> /* SPDX-License-Identifier: (GPL-2.0+ or MIT) */
>> /* SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) */
>>
>> So, should I fix the license as
>>
>> /* SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause) */
>>
>> just remove the '+' ?
> The simplest is to use the same license as your binding, so
> (GPL-2.0-only OR BSD-2-Clause) .

OK, I got it. Thank you very much.

>>>> +/*
>>>> + * Copyright (C) 2023 Nuvoton Technologies.
>>>> + * Author: Chi-Fen Li <cfli0@nuvoton.com>
>>>> + *
>>>> + * Device Tree binding constants for MA35D1 reset controller.
>>>> + */
>>>> +
>>>> +#ifndef __DT_BINDINGS_RESET_MA35D1_H
>>>> +#define __DT_BINDINGS_RESET_MA35D1_H
>>>> +
>>>> +#define MA35D1_RESET_CHIP	0
>>>> +#define MA35D1_RESET_CA35CR0	1
>>>> +#define MA35D1_RESET_CA35CR1	2
>>>> +#define MA35D1_RESET_CM4	3
>>>> +#define MA35D1_RESET_PDMA0	4
>>>> +#define MA35D1_RESET_PDMA1	5
>>>> +#define MA35D1_RESET_PDMA2	6
>>>> +#define MA35D1_RESET_PDMA3	7
>>>> +#define MA35D1_RESET_DISP	9
>>>> +#define MA35D1_RESET_VCAP0	10
>>>> +#define MA35D1_RESET_VCAP1	11
>>>> +#define MA35D1_RESET_GFX	12
>>>> +#define MA35D1_RESET_VDEC	13
>>>> +#define MA35D1_RESET_WHC0	14
>>>> +#define MA35D1_RESET_WHC1	15
>>>> +#define MA35D1_RESET_GMAC0	16
>>>> +#define MA35D1_RESET_GMAC1	17
>>>> +#define MA35D1_RESET_HWSEM	18
>>>> +#define MA35D1_RESET_EBI	19
>>>> +#define MA35D1_RESET_HSUSBH0	20
>>>> +#define MA35D1_RESET_HSUSBH1	21
>>>> +#define MA35D1_RESET_HSUSBD	22
>>>> +#define MA35D1_RESET_USBHL	23
>>>> +#define MA35D1_RESET_SDH0	24
>>>> +#define MA35D1_RESET_SDH1	25
>>>> +#define MA35D1_RESET_NAND	26
>>>> +#define MA35D1_RESET_GPIO	27
>>>> +#define MA35D1_RESET_MCTLP	28
>>>> +#define MA35D1_RESET_MCTLC	29
>>>> +#define MA35D1_RESET_DDRPUB	30
>>>> +#define MA35D1_RESET_TMR0	34
>>>> +#define MA35D1_RESET_TMR1	35
>>>> +#define MA35D1_RESET_TMR2	36
>>>> +#define MA35D1_RESET_TMR3	37
>>>> +#define MA35D1_RESET_I2C0	40
>>>> +#define MA35D1_RESET_I2C1	41
>>>> +#define MA35D1_RESET_I2C2	42
>>>> +#define MA35D1_RESET_I2C3	43
>>>> +#define MA35D1_RESET_QSPI0	44
>>>> +#define MA35D1_RESET_SPI0	45
>>>> +#define MA35D1_RESET_SPI1	46
>>>> +#define MA35D1_RESET_SPI2	47
>>>> +#define MA35D1_RESET_UART0	48
>>>> +#define MA35D1_RESET_UART1	49
>>>> +#define MA35D1_RESET_UART2	50
>>>> +#define MA35D1_RESET_UAER3	51
>>>> +#define MA35D1_RESET_UART4	52
>>>> +#define MA35D1_RESET_UART5	53
>>>> +#define MA35D1_RESET_UART6	54
>>>> +#define MA35D1_RESET_UART7	55
>>>> +#define MA35D1_RESET_CANFD0	56
>>>> +#define MA35D1_RESET_CANFD1	57
>>>> +#define MA35D1_RESET_EADC0	60
>>> Why do you have gaps here? These should be IDs, not register offsets.
>>>
>>> Best regards,
>>> Krzysztof
>>>
>> The reset controller registers are composed of four 32-bits register.
>> And the ID will be translated into the corresponding bit of these register.
>> Each ID is mapped to a unique bit position.
> That's not the usual way. IDs start from 0 or 1 and gets incremented by
> 1, without gaps. Remember that IDs cannot change and your register bits
> could have some changes (e.g. HW errata).
>
>
> Best regards,
> Krzysztof
>

OK, I will modify the ID as contiguous number, and create a lookup table 
to translate
ID into bitmap.


Best Regards,
Jacky Huang