linux-kernel - Re: [PATCH v7 8/8] iio: adc: ad4851: add ad485x driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <cpfpysi6zcr7n6qy642dmhsi2a4hom2p7l4rboj4jgakuzpa5p@5xaqd3oxxlrs>
Date: Sat, 7 Dec 2024 16:29:51 +0100
From: Uwe Kleine-König <u.kleine-koenig@...libre.com>
To: Antoniu Miclaus <antoniu.miclaus@...log.com>
Cc: jic23@...nel.org, robh@...nel.org, conor+dt@...nel.org, 
	dlechner@...libre.com, linux-iio@...r.kernel.org, devicetree@...r.kernel.org, 
	linux-kernel@...r.kernel.org, linux-pwm@...r.kernel.org
Subject: Re: [PATCH v7 8/8] iio: adc: ad4851: add ad485x driver

Hello Antoniu,

On Fri, Nov 29, 2024 at 05:35:46PM +0200, Antoniu Miclaus wrote:
> +static int ad4851_set_sampling_freq(struct ad4851_state *st, unsigned int freq)
> +{
> +	struct pwm_state cnv_state = {
> +		.duty_cycle = AD4851_T_CNVH_NS,
> +		.enabled = true,
> +	};
> +	int ret;
> +
> +	freq = clamp(freq, 1, st->info->max_sample_rate_hz);
> +
> +	cnv_state.period = DIV_ROUND_UP_ULL(NSEC_PER_SEC, freq);
> +
> +	ret = pwm_apply_might_sleep(st->cnv, &cnv_state);
> +	if (ret)
> +		return ret;

If you want to be sure that pwm_apply_might_sleep() doesn't round down
.period (and .duty_cycle?) too much, you might consider using
pwm_round_waveform_might_sleep(). But note that this function only works
for two hwpwm drivers currently.

> +
> +	st->sampling_freq = freq;
> +
> +	return 0;
> +}
> +
> +static const int ad4851_oversampling_ratios[] = {
> +	1, 2, 4, 8, 16,	32, 64, 128,
> +	256, 512, 1024, 2048, 4096, 8192, 16384, 32768,
> +	65536,
> +};
> +
> +static int ad4851_osr_to_regval(int ratio)

This function is called with an unsigned parameter only.

> +{
> +	int i;
> +
> +	for (i = 1; i < ARRAY_SIZE(ad4851_oversampling_ratios); i++)
> +		if (ratio == ad4851_oversampling_ratios[i])
> +			return i - 1;

Given that ad4851_oversampling_ratios[i] == 1 << i you could simplify
that. Something like:

	if (is_power_of_2(ratio) && ratio <= 65536 && ratio > 0)
		return ilog2(ratio);

> +
> +	return -EINVAL;
> +}
> +
> +static int ad4851_set_oversampling_ratio(struct ad4851_state *st,
> +					 const struct iio_chan_spec *chan,
> +					 unsigned int osr)
> +{
> +	unsigned int val;
> +	int ret;
> +
> +	guard(mutex)(&st->lock);
> +
> +	if (osr == 1) {
> +		ret = regmap_clear_bits(st->regmap, AD4851_REG_OVERSAMPLE,
> +					AD4851_OS_EN_MSK);
> +		if (ret)
> +			return ret;
> +
> +		ret = iio_backend_oversampling_disable(st->back);
> +		if (ret)
> +			return ret;
> +	} else {
> +		val = ad4851_osr_to_regval(osr);
> +		if (val < 0)
> +			return -EINVAL;
> +
> +		ret = regmap_set_bits(st->regmap, AD4851_REG_OVERSAMPLE,
> +				      AD4851_OS_EN_MSK);
> +		if (ret)
> +			return ret;
> +
> +		ret = regmap_update_bits(st->regmap, AD4851_REG_OVERSAMPLE,
> +					 AD4851_OS_RATIO_MSK, val);

You set this register twice in a row. Can this be done in a single
register access?

> +		if (ret)
> +			return ret;

Best regards
Uwe

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)