Devicetree
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Michael Reeves" <michael.reeves077@gmail.com>
Cc: devicetree@vger.kernel.org, robh@kernel.org, conor+dt@kernel.org
Subject: Re: [PATCH 4/4] arm64: dts: apple: t6031: Add CPU frequency scaling and OPPs
Date: Sat, 21 Nov 2026 11:53:20 +0000	[thread overview]
Message-ID: <sashiko-outbox-226994@kernel.org> (raw)
In-Reply-To: <20261121-m3-cpufreq-opp-v1-4-42a0c07c48a4@gmail.com>

Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [High] arm64: dts: apple: missing capacity-dmips-mhz on cpu_p14 and
  cpu_p15
- [High] arm64: dts: apple: missing cpufreq and capacity properties on
  t6032 die1 CPUs
--

commit bda66cc6e2f0fc13ab6cfd1d52b88625bed8a506
Author: Yureka Lilian <yureka@cyberchaos.dev>

arm64: dts: apple: t6031: Add CPU frequency scaling and OPPs

This commit adds cluster cpufreq controller nodes, operating performance
point tables, and CPU capacity values for the Apple T6031/T6034 Sawtooth
and Everest CPU clusters.

> diff --git a/arch/arm64/boot/dts/apple/t6031-base.dtsi b/arch/arm64/boot/dts/apple/t6031-base.dtsi
> index f62ce0690182b..772af1ea37b62 100644
> --- a/arch/arm64/boot/dts/apple/t6031-base.dtsi
> +++ b/arch/arm64/boot/dts/apple/t6031-base.dtsi
[ ... ]
> @@ -257,9 +296,12 @@ cpu_p13: cpu@10203 {
>  			cpu-release-addr = <0 0>; /* To be filled by loader */
>  			enable-method = "spin-table";
>  
> +			capacity-dmips-mhz = <1024>;
>  			d-cache-size = <0x20000>;
>  			i-cache-size = <0x30000>;
>  			next-level-cache = <&l2_cache_2>;
> +			operating-points-v2 = <&everest_opp>;
> +			performance-domains = <&cpufreq_p1>;
>  		};
>  
>  		cpu_p14: cpu@10204 {
> @@ -272,6 +314,8 @@ cpu_p14: cpu@10204 {
>  			d-cache-size = <0x20000>;
>  			i-cache-size = <0x30000>;
>  			next-level-cache = <&l2_cache_2>;
> +			operating-points-v2 = <&everest_opp>;
> +			performance-domains = <&cpufreq_p1>;
>  		};
>  
>  		cpu_p15: cpu@10205 {
> @@ -284,6 +328,8 @@ cpu_p15: cpu@10205 {
>  			d-cache-size = <0x20000>;
>  			i-cache-size = <0x30000>;
>  			next-level-cache = <&l2_cache_2>;
> +			operating-points-v2 = <&everest_opp>;
> +			performance-domains = <&cpufreq_p1>;
>  		};

[Severity: High]
Are cpu_p14 and cpu_p15 missing capacity-dmips-mhz = <1024>?

When topology_parse_cpu_capacity() sees capacity-dmips-mhz on earlier
CPUs and then encounters cpu_p14 without it, it logs a partial
information error, frees raw_capacity, and falls back to a uniform
capacity of 1024 for all CPUs, disabling asymmetric CPU capacity scaling
between the Sawtooth and Everest clusters on 16-core T6031 systems.

[ ... ]
> diff --git a/arch/arm64/boot/dts/apple/t6031-dieX.dtsi b/arch/arm64/boot/dts/apple/t6031-dieX.dtsi
> index 1624651c1d92c..bce4d677ae4c0 100644
> --- a/arch/arm64/boot/dts/apple/t6031-dieX.dtsi
> +++ b/arch/arm64/boot/dts/apple/t6031-dieX.dtsi
> @@ -7,6 +7,27 @@
>   */
>  
>  &SOC_NODE {
> +	DIE_NODE(cpufreq_e): performance-controller@210e20000 {
> +		compatible = "apple,t6031-cluster-cpufreq",
> +			     "apple,t8112-cluster-cpufreq";
> +		reg = <0x2 0x10e20000 0x0 0x1000>;
> +		#performance-domain-cells = <0>;
> +	};
> +
> +	DIE_NODE(cpufreq_p0): performance-controller@211e20000 {
> +		compatible = "apple,t6031-cluster-cpufreq",
> +			     "apple,t8112-cluster-cpufreq";
> +		reg = <0x2 0x11e20000 0x0 0x1000>;
> +		#performance-domain-cells = <0>;
> +	};
> +
> +	DIE_NODE(cpufreq_p1): performance-controller@212e20000 {
> +		compatible = "apple,t6031-cluster-cpufreq",
> +			     "apple,t8112-cluster-cpufreq";
> +		reg = <0x2 0x12e20000 0x0 0x1000>;
> +		#performance-domain-cells = <0>;
> +	};

[Severity: High]
Should the die 1 CPU nodes (cpu_e10..cpu_p35) in t6032.dtsi also be
updated with capacity-dmips-mhz, operating-points-v2, and
performance-domains?

t6032.dtsi includes t6031-base.dtsi and includes t6031-dieX.dtsi for both
die 0 and die 1, which instantiates cpufreq_{e,p0,p1}_die1.  Leaving the
die 1 CPUs in t6032.dtsi without these properties causes
topology_parse_cpu_capacity() to abort capacity scaling across all CPUs on
T6032, causes apple_soc_cpufreq_init() to fail with -ENOENT on all die 1
CPUs, and prevents init_cpu_capacity_callback() from ever emptying
cpus_to_visit.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261121-m3-cpufreq-opp-v1-0-42a0c07c48a4@gmail.com?part=4

  parent reply	other threads:[~2026-11-21 11:53 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-21 11:42 [PATCH 0/4] arm64: dts: apple: Add CPU frequency scaling and OPPs for M3 family Michael Reeves via B4 Relay
2026-11-21 11:42 ` [PATCH 1/4] dt-bindings: cpufreq: apple,cluster-cpufreq: Add M3 family compatibles Michael Reeves via B4 Relay
2026-11-21 11:42 ` [PATCH 2/4] arm64: dts: apple: t8122: Add CPU frequency scaling and OPPs Michael Reeves via B4 Relay
2026-11-21 11:42 ` [PATCH 3/4] arm64: dts: apple: t6030: " Michael Reeves via B4 Relay
2026-11-21 11:42 ` [PATCH 4/4] arm64: dts: apple: t6031: " Michael Reeves via B4 Relay
2026-11-21 11:52   ` Michael Reeves
2026-11-21 12:11     ` Michael Reeves
2026-11-21 11:53   ` sashiko-bot [this message]
2026-11-21 11:52 ` [PATCH 0/4] arm64: dts: apple: Add CPU frequency scaling and OPPs for M3 family Janne Grunau

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-226994@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=conor+dt@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=michael.reeves077@gmail.com \
    --cc=robh@kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox