Re: [PATCH v8 1/2] bus: simple-pm-bus: Remove child devices when the bus is unbound

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The bus device is left in the RPM_ACTIVE state with its clocks permanently enabled upon removal, as `pm_runtime_disable()` cancels the asynchronous idle request triggered by child device removal.
--

commit db0675ba90c5bdf94a4790c06b36a7b5f8dff5e4
Author: Herve Codina <herve.codina@bootlin.com>

bus: simple-pm-bus: Remove child devices when the bus is unbound

During probe, the simple-pm-bus driver populates child devices but does not
remove them during its removal. This commit explicitly depopulates child
devices when the bus is unbound to keep the device state consistent.

> --- a/drivers/bus/simple-pm-bus.c
> +++ b/drivers/bus/simple-pm-bus.c
> @@ -83,6 +83,9 @@ static void simple_pm_bus_remove(struct platform_device *pdev)
>  
>  	dev_dbg(&pdev->dev, "%s\n", __func__);
>  
> +	if (pdev->dev.of_node)
> +		of_platform_depopulate(&pdev->dev);
> +
>  	pm_runtime_disable(&pdev->dev);

[Severity: High]
Will this sequence leave the bus device active and leak clock enable counts?

When of_platform_depopulate() removes the child devices, the parent bus
device's child_count drops to zero, triggering an asynchronous
pm_request_idle() on the parent.

Because pm_runtime_disable() is called immediately afterwards, it will invoke
__pm_runtime_barrier() and synchronously cancel any pending asynchronous PM
requests.

If the idle request is cancelled before it executes, does the bus bypass
simple_pm_bus_runtime_suspend() and remain in the RPM_ACTIVE state with its
clocks permanently enabled? When devres later frees the clock structures via
clk_bulk_put_all(), would the active clock enable counts be leaked, causing
a warning? Should there be a synchronous runtime PM put or suspend operation
before disabling runtime PM?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260630102717.413501-1-herve.codina@bootlin.com?part=1