linux-kernel - Re: [PATCH 1/3 v2] regulator: act8865: add PMIC act8865 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: <20131217130653.GA3185@sirena.org.uk>
Date:	Tue, 17 Dec 2013 13:06:53 +0000
From:	Mark Brown <broonie@...nel.org>
To:	Wenyou Yang <wenyou.yang@...el.com>
Cc:	lgirdwood@...il.com, linux-kernel@...r.kernel.org,
	grant.likely@...aro.org, rob.herring@...xeda.com,
	linux-doc@...r.kernel.org, vpalatin@...omium.org,
	nicolas.ferre@...el.com, plagnioj@...osoft.com,
	linux-arm-kernel@...ts.infradead.org, devicetree@...r.kernel.org
Subject: Re: [PATCH 1/3 v2] regulator: act8865: add PMIC act8865 driver

On Tue, Dec 17, 2013 at 01:36:35PM +0800, Wenyou Yang wrote:
>  Signed-off-by: Wenyou Yang <wenyou.yang@...el.com>

Much better.  Still a few small issues though.

> +/* ACt8865 voltage table */
> +static const u32 act8865_voltages_table[] = {

This is the wrong type for a voltage table but it looks like it
shouldn't be a voltage table at all - this looks like it should be
mapped as linear ranges.

> +static struct regulator_ops act8865_ops = {
> +	.list_voltage		= regulator_list_voltage_table,
> +	.get_voltage_sel	= regulator_get_voltage_sel_regmap,
> +	.set_voltage_sel	= regulator_set_voltage_sel_regmap,
> +	.enable			= regulator_enable_regmap,
> +	.disable		= regulator_disable_regmap,
> +	.is_enabled		= regulator_is_enabled_regmap,
> +	.set_suspend_voltage	= act8865_set_suspend_voltage,
> +	.set_suspend_enable	= regulator_enable_regmap,
> +	.set_suspend_disable	= regulator_disable_regmap,
> +};

This is missing a map_voltage() operation.

> +#ifdef CONFIG_OF
> +static const struct of_device_id act8865_dt_ids[] = {
> +	{ .compatible = "active-semi,act8865" },
> +	{ }
> +};
> +MODULE_DEVICE_TABLE(of, act8865_dt_ids);

active-semi needs adding to vendor-prefixes.txt.

> +	np = of_find_node_by_name(dev->of_node, "regulators");
> +	if (!np) {
> +		dev_err(dev, "missing 'regulators' subnode in DT\n");
> +		return -EINVAL;
> +	}
> +
> +	for (i = 0; i < ARRAY_SIZE(rmatch); i++)
> +		rmatch[i].name = act8865_reg[i].name;
> +
> +	matched = of_regulator_match(dev, np, rmatch, ARRAY_SIZE(rmatch));
> +	if (matched <= 0)
> +		return matched;

You should register all regulators the device has regardless of how many
have parameters configured in the DT so that the settings can be read
back for diagnostic purposes.

> +	if (dev->of_node && !pdata) {
> +		const struct of_device_id *id;
> +		struct act8865_platform_data pdata_of;
> +
> +		id = of_match_device(of_match_ptr(act8865_dt_ids), dev);
> +		if (!id)
> +			return -ENODEV;
> +
> +		ret = act8865_pdata_from_dt(dev, of_node, &pdata_of);
> +		if (ret < 0)
> +			return ret;
> +
> +		pdata = &pdata_of;
> +	} else {
> +		memset(of_node, 0, sizeof(of_node));
> +	}

What's this memset() for?

> +		rdev[i] = devm_regulator_register(&client->dev,
> +						&act8865_reg[i], &config);

> +err_unregister:
> +	while (--i >= 0)
> +		regulator_unregister(rdev[i]);

The whole purpose of devm_ APIs is to avoid the need to explicitly
unregister, you shouldn't need to call regulator_unregister() at all.

Download attachment "signature.asc" of type "application/pgp-signature" (837 bytes)