Re: [PATCH v10 2/4] iio: pressure: dps310: derive the drain interval from the watermark

[Andy Shevchenko]

On Wed, Sep 30, 2026 at 01:01:56PM +0300, Rupesh Majhi wrote:
> The drain interval only tracked how fast the FIFO fills, so a small
> watermark still waited half a FIFO for its samples. Take the watermark
> as the number of scans to wait for, but never drain slower than half a
> fill.

...

> -	unsigned int fill_ms;
> +	unsigned int fill_ms, want_ms;

What about

	unsigned int drain_interval_ms, fill_ms, want_ms;

> +	int scan_rate;

>  	fill_ms = MSEC_PER_SEC * DPS310_FIFO_DEPTH / (prs_rate + tmp_rate);
>  
> -	return clamp(fill_ms / 2, DPS310_DRAIN_MIN_MS, DPS310_DRAIN_MAX_MS);
> +	/*
> +	 * There is no hardware watermark, so take it as the number of scans
> +	 * the user will wait for and drain at that rate. Scans come from
> +	 * whichever measurement drives them.
> +	 */
> +	if (test_bit(DPS310_SCAN_PRESSURE, data->iio->active_scan_mask))
> +		scan_rate = prs_rate;
> +	else
> +		scan_rate = tmp_rate;
> +
> +	want_ms = data->watermark * MSEC_PER_SEC / scan_rate;

> +	return clamp(min(want_ms, fill_ms / 2), DPS310_DRAIN_MIN_MS,
> +		     DPS310_DRAIN_MAX_MS);

	drain_interval_ms = min(want_ms, fill_ms / 2);

	return clamp(drain_interval_ms, DPS310_DRAIN_MIN_MS, DPS310_DRAIN_MAX_MS);

?

-- 
With Best Regards,
Andy Shevchenko