Re: [PATCH 6/9] iio: adc: rzt2h: expose sampling frequency

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Spurious -ETIMEDOUT on single reads due to extremely short 1-jiffy completion timeout
--

commit c401ad2e533dc2bc5a1b011d098327c8d7400f4f
Author: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>

iio: adc: rzt2h: expose sampling frequency
    
This patch exposes the sampling frequency as a per-channel
IIO_CHAN_INFO_SAMP_FREQ attribute to allow userspace control over conversion
time. It derives the frequency from the ADCLK rate and programs the sample
time into ADSSTRn for each enabled channel.

> diff --git a/drivers/iio/adc/rzt2h_adc.c b/drivers/iio/adc/rzt2h_adc.c
> index d76226375f220..2e13f931a2e9a 100644
> --- a/drivers/iio/adc/rzt2h_adc.c
> +++ b/drivers/iio/adc/rzt2h_adc.c

[ ... ]

> @@ -124,6 +146,8 @@ static int rzt2h_adc_read_single(struct rzt2h_adc *adc, unsigned int ch, int *va
>  	/* Enable a single channel */
>  	writew(RZT2H_ADANSA0_CH_MASK(ch), adc->base + RZT2H_ADANSA0_REG);
>  
> +	rzt2h_adc_set_sst(adc, ch, adc->sst[ch]);
> +
>  	rzt2h_adc_start(adc, RZT2H_ADCSR_ADCS_SINGLE);
>  
>  	/*

[Severity: High]
This is a pre-existing issue, but does the newly introduced configurable
sampling frequency exacerbate a spurious timeout bug here?

In rzt2h_adc_read_single(), right below this hunk, the code calls:

	ret = wait_for_completion_timeout(&adc->completion, usecs_to_jiffies(1));

Because usecs_to_jiffies(1) evaluates to 1, the timeout is set to exactly
1 jiffy, which will expire at the very next timer tick.

If the tick fires immediately after the wait starts, it aborts almost
instantly (returning 0) before the hardware interrupt arrives, causing the
kernel to intermittently and incorrectly return -ETIMEDOUT to userspace
instead of the actual ADC reading.

Since the patch now allows increasing the hardware conversion delay (up to
268 cycles or ~5.36us at 50MHz ADCLK), does this widen the race window and
significantly increase the probability of this spurious timeout occurring?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260828145943.2077589-1-cosmin-gabriel.tanislav.xa@renesas.com?part=6