Re: [PATCH 10/10] arm64: dts: apple: Add minimal t8132 (M4) device trees

On 7/5/26 15:28, [email protected] wrote:
> Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
> - [Medium] The `aic` node has a mismatched unit address and overlapping memory regions in its `reg` property.
> - [Low] The `model` property for the MacBook Air (13-inch) erroneously declares the year as 2024 instead of 2025.
> - [Low] `t8132` device trees are incorrectly placed at the very top of the `Makefile`, breaking the chronological sorting order.
> --
>
> commit 281d92c4ca5eb8a2224c6bb0b9116c7518b422d1
> Author: Yureka Lilian <[email protected]>
>
> arm64: dts: apple: Add minimal t8132 (M4) device trees
>
> This commit adds minimal device trees for a series of new Apple M4 devices,
> including MacBook Pros, MacBook Airs, iMacs, and the Mac mini. The device
> trees are limited to essential components like CPU cores, interrupts, and
> serial, allowing a basic boot. It also notes that SMP support currently
> remains unreliable.
>
>> diff --git a/arch/arm64/boot/dts/apple/Makefile b/arch/arm64/boot/dts/apple/Makefile
>> index 6fc3349a58428..181bd0e289653 100644
>> --- a/arch/arm64/boot/dts/apple/Makefile
>> +++ b/arch/arm64/boot/dts/apple/Makefile
>> @@ -1,4 +1,10 @@
>>   # SPDX-License-Identifier: GPL-2.0
>> +dtb-$(CONFIG_ARCH_APPLE) += t8132-j604.dtb
>> +dtb-$(CONFIG_ARCH_APPLE) += t8132-j623.dtb
>> +dtb-$(CONFIG_ARCH_APPLE) += t8132-j624.dtb
>> +dtb-$(CONFIG_ARCH_APPLE) += t8132-j713.dtb
>> +dtb-$(CONFIG_ARCH_APPLE) += t8132-j715.dtb
>> +dtb-$(CONFIG_ARCH_APPLE) += t8132-j773g.dtb
>>   dtb-$(CONFIG_ARCH_APPLE) += s5l8960x-j71.dtb
> [Severity: Low]
> This isn't a bug, but should these new t8132 device trees be appended
> to the bottom of the list rather than prepended at the top, to maintain
> chronological sorting order in the Makefile?
I had this at the top to make rebasing easier, but for the upstream 
version it should be at the bottom. Changed locally.
>
> [ ... ]
>
>> diff --git a/arch/arm64/boot/dts/apple/t8132-j713.dts b/arch/arm64/boot/dts/apple/t8132-j713.dts
>> index 0000000000000..b1f779c37d5fe
>> --- /dev/null
>> +++ b/arch/arm64/boot/dts/apple/t8132-j713.dts
> [ ... ]
>> +/ {
>> +	compatible = "apple,j713", "apple,t8132", "apple,arm-platform";
>> +	model = "Apple MacBook Air (13-inch, M4, 2024)";
> [Severity: Low]
> This isn't a bug, but does the model string contain a typo? The header
> comment for this file identifies the target as a 2025 model, but the model
> string exposes 2024 to userspace.
Typo. Fixed for v2.
>
> [ ... ]
>
>> diff --git a/arch/arm64/boot/dts/apple/t8132.dtsi b/arch/arm64/boot/dts/apple/t8132.dtsi
>> index 0000000000000..715980b5c7de4
>> --- /dev/null
>> +++ b/arch/arm64/boot/dts/apple/t8132.dtsi
> [ ... ]
>> +		aic: interrupt-controller@381180000 {
>> +			compatible = "apple,t8132-aic3", "apple,t8122-aic3";
>> +			#interrupt-cells = <3>;
>> +			interrupt-controller;
>> +			reg = <0x00000003 0x81000000 0x00000000 0x1cc000>,
>> +			      <0x00000003 0x81040000 0x00000000 0x4000>;
> [Severity: Medium]
> Is the base address and memory region mapping for aic intentional?
>
> The node defines a unit address of @381180000, but the reg property specifies
> a base of 0x81000000 (translating to 0x381000000). Also, the first memory
> region of size 0x1cc000 completely overlaps the second region starting at
> 0x81040000.
>
> Could this typo in the physical address map explain the unreliable smp boot
> behavior mentioned in the commit message?

I believe the core and event regs overlapping is correct. The same 
overlap exists in t8122, and also in downstream Asahi t6030/t6031 dts.

>