[PATCH 1/8] arm64: dts: s32g2: Add the STM description
Daniel Lezcano
daniel.lezcano at linaro.org
Fri Aug 1 01:35:04 PDT 2025
On 01/08/2025 01:20, Frank Li wrote:
> On Wed, Jul 30, 2025 at 11:15:40PM +0200, Daniel Lezcano wrote:
[ ... ]
>>>> + compatible = "nxp,s32g2-stm";
>>>> + reg = <0x4011c000 0x3000>;
>>>> + interrupts = <GIC_SPI 24 IRQ_TYPE_LEVEL_HIGH>;
>>>> + clocks = <&clks 0x3b>, <&clks 0x3c>, <&clks 0x3c>;
>>>> + clock-names = "counter", "module", "register";
>>>> + status = "disabled";
>>>
>>> why not default enable.
>>
>> The S32G2 and S32G3 can have different variants with 2, 4, 8 Cortex-A53 and
>> 3 or 4 Cortex-M7. We enable the same number of CPUs present on the system.
>>
>> AFAIU:
>> S32G233A : 2 x Cortex-A53
>> S32G274A : 4 x Cortex-A53
>>
>> S32G399A : 8 x Cortex-A53
>> S32G379A : 4 x Cortex-A53
>>
>> Otherwise we would have to do the opposite, that is disable the unused ones
>> in the s32g274a-rdb2.dts, s32g399a-rdb3.dts and other dts which include the
>> s32g2.dtsi and s32g3.dtsi.
>>
>
> That's fine by default disabled. but what's impact if it is enable but no
> one use it?
At the first glance I would say we call the probe function for nothing,
so adding an extra overhead. When repeated into multiple drivers that
increases the boot time significantly. It is certainly a good practice
as a rule of thumb to enable only the ones we really need.
--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog
More information about the linux-arm-kernel
mailing list