Re: [PATCH 5/7] spmi: apple: lock around FIFOs

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Using an interruptible mutex for a fast hardware bus lock introduces unexpected I/O failures.
--

--- Patch [5]: [PATCH 5/7] spmi: apple: lock around FIFOs ---
commit c446ad8ec5411cb7c1e85b2172699c9fc3a53edd
Author: Alba Mendez <me@alba.sh>

spmi: apple: lock around FIFOs

The driver was missing locking around register interactions.

> diff --git a/drivers/spmi/spmi-apple-controller.c b/drivers/spmi/spmi-apple-controller.c
> index d69d16665ce85..d8509f027df69 100644
> --- a/drivers/spmi/spmi-apple-controller.c
> +++ b/drivers/spmi/spmi-apple-controller.c
[ ... ]
> @@ -73,6 +75,10 @@ static int spmi_raw_cmd(struct spmi_controller *ctrl, u8 opc, u8 sid,
>  	size_t i = 0, j;
>  	int ret;
>  
> +	ret = mutex_lock_interruptible(&spmi->fifo_lock);
> +	if (ret)
> +		return ret;

[Severity: High]
Can using an interruptible lock here cause unexpected I/O failures?

If an SPMI read or write blocks on this lock while another transaction is 
ongoing, and the blocking process receives a signal, the lock acquisition 
fails and returns -EINTR.

Higher-level frameworks like regulators or PMIC subsystems generally do not 
expect or retry hardware bus operations failing with -EINTR. This could 
cause critical operations, such as CPU voltage changes during cpufreq scaling, 
to fail randomly if a signal is pending.

Would it be safer to use a standard uninterruptible mutex_lock() to serialize 
this fast hardware bus I/O and prevent these systemic side effects?

> +
>  	writel(spmi_cmd, spmi->regs + SPMI_CMD_REG);
>  
>  	while (i < len) {

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260725-t603x-spmi-v1-0-e1a29fcd2d38@chaosmail.tech?part=5