On 05/05/15 16:42, Philipp Zabel wrote:
Am Dienstag, den 05.05.2015, 17:19 +0200 schrieb Maxime Coquelin:
For example, includes/dt-bindings/mfd/stm32f4-rcc.h would look like:
#define GPIOA 0
#define GPIOB 1
...
#define LTDC 186
That looks a bit fragile.
At least the defines for the indices should be properly namespaced,
check out include/dt-bindings/gpio/tegra-gpio.h for a similar case.
Good point.
#define STM32F4_RESET(x) (x + 128)
#define STM32F4_CLOCK(x) (x + 384)
Thinking more about this point, if we are going to follow hardware if
might be better to have:
#define STM32F4_RCC_AHB1_GPIOA 0
#define STM32F4_RCC_AHB1_GPIOA 1
...
#define STM32F4_RCC_APB2_LTDC 26
#define STM32F4_AHB1_RESET(x) (STM32F4_RCC_AHB1_##x##_BIT + (0x10 * 8))
#define STM32F4_AHB2_RESET(x) (STM32F4_RCC_AHB2_##x##_BIT + (0x14 * 8))
...
#define STM32F4_APB2_RESET(x) (STM32F4_RCC_APB2_##x##_BIT + (0x24 * 8))
Its more typing (or copy 'n pasting) by at least every number now maps
directly to the datasheet.
Then, in DT, a reset would be described like this:
timer2 {
resets = <&rcc STM32F4_RESET(TIM2)>;
};
Phillip, Daniel, does that look acceptable to you?
Doesn't look unreasonable.
I am a little uneasy simply because there are very few similar header files
in that directory but I haven't thought of a better idea.
Since this file will be shared by both clock and reset drivers, I
don't see better option.
I will implement it in v8 if Philipp agrees.
Are the device tree maintainers happy with this idiom spreading?
Except for the point above, I think this is acceptable.
regards
Philipp
--
To unsubscribe from this list: send the line "unsubscribe linux-arch" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
