Devicetree
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Sreeshankar K" <sreeshankar0910@gmail.com>
Cc: linux-clk@vger.kernel.org, devicetree@vger.kernel.org,
	robh@kernel.org, conor+dt@kernel.org
Subject: Re: [PATCH v15 8/8] arm64: dts: qcom: Add SM7250 devicetree file
Date: Mon, 16 Nov 2026 06:16:22 +0000	[thread overview]
Message-ID: <sashiko-outbox-216254@kernel.org> (raw)
In-Reply-To: <20261116-sm7250-v15-8-6acbb232fef8@gmail.com>

Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [Medium] cpufreq: qcom-hw: partial init and resource leak in
  qcom_cpufreq_hw_cpu_init()
- [Low] arm64: dts: qcom: sm7250: oversized PDC register region in pdc
  node
--

commit a2793b66bb2274241a395fec222892481d4a324f
Author: Sreeshankar K <sreeshankar0910@gmail.com>

arm64: dts: qcom: Add SM7250 devicetree file

This commit adds the base devicetree file for the Qualcomm SM7250 SoC,
including CPU, PSCI, clock, interrupt, and serial console nodes.

> diff --git a/arch/arm64/boot/dts/qcom/sm7250.dtsi b/arch/arm64/boot/dts/qcom/sm7250.dtsi
> new file mode 100644
> index 0000000000000..231b6ea59a590
> --- /dev/null
> +++ b/arch/arm64/boot/dts/qcom/sm7250.dtsi
[ ... ]
> +		pdc: interrupt-controller@b220000 {
> +			compatible = "qcom,sm7250-pdc", "qcom,pdc";
> +			reg = <0 0xb220000 0 0x30000>, <0 0x17c000f0 0 0x60>;

[Severity: Low]
Should the first reg region size be 0x10000 instead of 0x30000?

qcom_pdc_probe() maps max_t(resource_size_t, resource_size(&res),
PDC_DRV_SIZE) where PDC_DRV_SIZE is 0x10000, so specifying 0x30000 maps
three PDC DRV regions instead of just the OS PDC DRV region.

[ ... ]

> +		cpufreq_hw: cpufreq@18323000 {
> +			compatible = "qcom,sm7250-cpufreq-hw", "qcom,cpufreq-hw";
> +			reg = <0 0x18323000 0 0x1000>,
> +			      <0 0x18325800 0 0x1000>,
> +			      <0 0x18327800 0 0x1000>;
> +			reg-names = "freq-domain0",
> +				    "freq-domain1",
> +				    "freq-domain2";
> +			interrupts-extended = <&lmh_cluster0 0>,
> +					      <&lmh_cluster1 0>;
> +			interrupt-names = "dcvsh-irq-0",
> +					  "dcvsh-irq-1";

[Severity: Medium]
Can defining two LMh interrupts across three frequency domains cause
partial cpufreq initialization, leaked frequency tables, and dangling
data->policy pointers?

If qcom_cpufreq_hw_driver_probe() runs when lmh_cluster0 and lmh_cluster1
have not registered their IRQ domains (for example with
CONFIG_ARM_QCOM_CPUFREQ_HW=y and CONFIG_QCOM_LMH=n, or before qcom-lmh.ko
loads), platform_get_irq_optional(pdev, index) in
qcom_cpufreq_hw_lmh_init() fails for domains 0 and 1 (CPUs 0..6) but
returns -ENXIO for domain 2 (CPU 7), which qcom_cpufreq_hw_lmh_init()
treats as success:

drivers/cpufreq/qcom-cpufreq-hw.c:qcom_cpufreq_hw_cpu_init() {
    ...
	data->policy = policy;

	ret = qcom_cpufreq_hw_read_lut(cpu_dev, policy);
	if (ret) {
		dev_err(dev, "Domain-%d failed to read LUT\n", index);
		return ret;
	}

	ret = dev_pm_opp_get_opp_count(cpu_dev);
	if (ret <= 0) {
		dev_err(cpu_dev, "Failed to add OPPs\n");
		return -ENODEV;
	}

	return qcom_cpufreq_hw_lmh_init(policy, index);
}

When qcom_cpufreq_hw_lmh_init() fails on CPUs 0..6, cpufreq_online()
skips cpufreq_driver->exit() and frees policy, leaking policy->freq_table
and dynamic OPPs while leaving data->policy pointing to the freed policy
(which qcom_cpufreq_hw_recalc_rate() later dereferences).  Because CPU 7
succeeds, cpufreq_register_driver() sees a non-empty cpufreq_policy_list
and returns 0 instead of deferring probe, leaving CPUs 0..6 without
cpufreq support.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261116-sm7250-v15-0-6acbb232fef8@gmail.com?part=8

      reply	other threads:[~2026-11-16  6:16 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-16  5:33 [PATCH v15 0/8] Add initial support for Qualcomm SM7250 SoC Sreeshankar K
2026-11-16  5:33 ` [PATCH v15 1/8] dt-bindings: clock: Add SM7250 global clock controller Sreeshankar K
2026-11-16  5:33 ` [PATCH v15 2/8] clk: qcom: Add Global Clock Controller (GCC) driver for SM7250 Sreeshankar K
2026-11-16  5:33 ` [PATCH v15 3/8] thermal/qcom/lmh: Switch to match data for SoC CPU-to-node mappings Sreeshankar K
2026-11-16  5:43   ` sashiko-bot
2026-11-16  5:33 ` [PATCH v15 4/8] dt-bindings: thermal: lmh: Document SM7250 LMh Sreeshankar K
2026-11-16  5:33 ` [PATCH v15 5/8] thermal/qcom/lmh: Add support for Qualcomm SM7250 Sreeshankar K
2026-11-16  5:33 ` [PATCH v15 6/8] dt-bindings: firmware: qcom,scm: Document SM7250 SCM Sreeshankar K
2026-11-16  5:33 ` [PATCH v15 7/8] dt-bindings: cpufreq: qcom-hw: Document SM7250 cpufreq hardware Sreeshankar K
2026-11-16  5:33 ` [PATCH v15 8/8] arm64: dts: qcom: Add SM7250 devicetree file Sreeshankar K
2026-11-16  6:16   ` sashiko-bot [this message]

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-216254@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=conor+dt@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=linux-clk@vger.kernel.org \
    --cc=robh@kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=sreeshankar0910@gmail.com \
    /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