Re: [PATCH v9 5/9] mfd: mt6397: Add support for MT6392 PMIC

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:

Pre-existing issues:
- [High] Resource leaks and Use-After-Free due to missing PM notifier teardown
- [High] Race condition in PM notifier causing lost IRQ hardware state
--

commit 4936b4ce67a0024573fa33154f81b91a5d2261b2
Author: Fabien Parent <parent.f@gmail.com>

mfd: mt6397: Add support for MT6392 PMIC

Align the MT6397 PMIC driver to other MFD drivers by passing only an
identifier through mt6397_of_match[*].data and add support for the MT6392
PMIC and its regulator, RTC, keys and pinctrl devices.

> diff --git a/drivers/mfd/mt6397-irq.c b/drivers/mfd/mt6397-irq.c
> index 5d2e5459f7444..80ea5b92d232f 100644
> --- a/drivers/mfd/mt6397-irq.c
> +++ b/drivers/mfd/mt6397-irq.c
[ ... ]
> @@ -203,6 +205,12 @@ int mt6397_irq_init(struct mt6397_chip *chip)
>  		chip->int_status[0] = MT6397_INT_STATUS0;
>  		chip->int_status[1] = MT6397_INT_STATUS1;
>  		break;
> +	case MT6392_CHIP_ID:
> +		chip->int_con[0] = MT6392_INT_CON0;
> +		chip->int_con[1] = MT6392_INT_CON1;
> +		chip->int_status[0] = MT6392_INT_STATUS0;
> +		chip->int_status[1] = MT6392_INT_STATUS1;
> +		break;
>  
>  	default:
>  		dev_err(chip->dev, "unsupported chip: 0x%x\n", chip->chip_id);

[Severity: High]
This isn't a bug introduced by this patch, but does the global PM notifier
leak if probe fails?

The PM notifier is registered globally at the end of mt6397_irq_init() without
using a devm_action wrapper:

	register_pm_notifier(&chip->pm_nb);
	return 0;

If probe fails after IRQ initialization (for example, if
devm_mfd_add_devices() fails in mt6397_probe()), or if the driver is unbound
via sysfs, devres frees the pmic structure while the PM notifier remains
active in the global kernel chain.

Could the next system suspend/resume cycle then call the notifier callback
on freed memory, leading to a crash?

Additionally, does the irq domain leak on unbind since mt6397_driver lacks a
.remove callback entirely?

[Severity: High]
This is a pre-existing issue, but can the PM notifier corrupt hardware state?

The PM notifier callback (mt6397_irq_pm_notifier) writes to the int_con
hardware registers using cached masks without acquiring chip->irqlock:

	case PM_SUSPEND_PREPARE:
		regmap_write(chip->regmap,
			     chip->int_con[0], chip->wake_mask[0]);

However, core IRQ functions like mt6397_irq_sync_unlock() modify the exact
same int_con registers safely under the protection of irqlock:

	regmap_write(mt6397->regmap, mt6397->int_con[0],
		     mt6397->irq_masks_cur[0]);

Since user tasks are still running during PM_SUSPEND_PREPARE and are already
thawed prior to PM_POST_SUSPEND, concurrent calls to enable_irq(),
disable_irq(), or irq_set_wake() can interleave register reads and writes.

Could this cause the hardware to be programmed with stale interrupt masks or
overwrite the suspend wake mask, leading to dropped device interrupts or a
failure to wake the system?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260621081634.467858-1-l.scorcia@gmail.com?part=5