| 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
| ||
|
Message-ID: <20240702180100.00000096@Huawei.com>
Date: Tue, 2 Jul 2024 18:01:00 +0100
From: Jonathan Cameron <Jonathan.Cameron@...wei.com>
To: Thomas Bonnefille <thomas.bonnefille@...tlin.com>
CC: Jonathan Cameron <jic23@...nel.org>, Lars-Peter Clausen <lars@...afoo.de>,
Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>,
Conor Dooley <conor+dt@...nel.org>, Chen Wang <unicorn_wang@...look.com>,
Inochi Amaoto <inochiama@...look.com>, "Paul Walmsley"
<paul.walmsley@...ive.com>, Palmer Dabbelt <palmer@...belt.com>, Albert Ou
<aou@...s.berkeley.edu>, Thomas Petazzoni <thomas.petazzoni@...tlin.com>,
Miquèl Raynal <miquel.raynal@...tlin.com>,
<linux-iio@...r.kernel.org>, <devicetree@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, <linux-riscv@...ts.infradead.org>
Subject: Re: [PATCH 2/3] iio: adc: sophgo-saradc: Add driver for Sophgo
SARADC
On Tue, 02 Jul 2024 13:52:22 +0200
Thomas Bonnefille <thomas.bonnefille@...tlin.com> wrote:
> This adds a driver for the common Sophgo SARADC.
>
> Signed-off-by: Thomas Bonnefille <thomas.bonnefille@...tlin.com>
Hi Thomas,
Welcome to IIO.
Various comments inline
Thanks,
Jonathan
> diff --git a/drivers/iio/adc/sophgo-adc.c b/drivers/iio/adc/sophgo-adc.c
> new file mode 100644
> index 000000000000..a94d839d40ec
> --- /dev/null
> +++ b/drivers/iio/adc/sophgo-adc.c
> @@ -0,0 +1,223 @@
> +// SPDX-License-Identifier: GPL-2.0-or-later
> +/*
> + * Sophgo SARADC Driver
> + *
> + * Copyright (C) Bootlin 2024
> + * Author: Thomas Bonnefille <thomas.bonnefille@...tlin.com>
> + */
> +
> +#include <linux/clk.h>
> +#include <linux/completion.h>
> +#include <linux/dev_printk.h>
> +#include <linux/interrupt.h>
> +#include <linux/iio/iio.h>
> +#include <linux/iopoll.h>
> +#include <linux/module.h>
> +#include <linux/mutex.h>
> +#include <linux/of.h>
probably want mod_devicetable.h instead.
> +#include <linux/platform_device.h>
> +
> +#define SOPHGO_SARADC_CTRL_REG 0x04
> +#define SOPHGO_SARADC_EN BIT(0)
> +#define SOPHGO_SARADC_SEL(x) BIT((x)+4)
> +#define SOPHGO_SARADC_STATUS_REG 0x08
> +#define SOPHGO_SARADC_BUSY BIT(0)
> +#define SOPHGO_SARADC_CYC_SET_REG 0x0C
> +#define SOPHGO_SARADC_DEF_CYC_SETTINGS 0xF1F0F
> +#define SOPHGO_SARADC_CH_RESULT_REG(x) (0x10+4*(x))
> +#define SARADC_CH_RESULT(x) ((x) & 0xfff)
> +#define SARADC_CH_VALID(x) ((x) & BIT(15))
Prefer defined masks and FIELD_GET() to actually get the values.
Also use consistent prefix for defines.
> +#define SOPHGO_SARADC_INTR_EN_REG 0x20
> +#define SOPHGO_SARADC_INTR_CLR_REG 0x24
> +
> +#define SOPHGO_SARADC_CHANNEL(index) \
> + { \
> + .type = IIO_VOLTAGE, \
> + .indexed = 1, \
> + .channel = index, \
> + .info_mask_separate = BIT(IIO_CHAN_INFO_RAW), \
> + .info_mask_shared_by_type = BIT(IIO_CHAN_INFO_SCALE), \
> + .scan_index = index, \
> + .scan_type = { \
> + .sign = 'u', \
> + .realbits = 12, \
No need to define these if not providing buffered capture.
> + }, \
> + }
> +
> +struct sophgo_saradc {
> + struct completion completion;
> + struct iio_info info;
> + void __iomem *regs;
> + struct mutex lock;
Add a comment on what data this lock protects.
> + struct clk *clk;
After changes below, probably don't need to hang on to the clk.
> + int irq;
Rarely need to keep hold of the irq after probe. Use a local variable for
it in instead.
> +};
> +
> +static const struct iio_chan_spec sophgo_channels[] = {
> + SOPHGO_SARADC_CHANNEL(1),
> + SOPHGO_SARADC_CHANNEL(2),
> + SOPHGO_SARADC_CHANNEL(3),
> +};
> +
> +static void sophgo_saradc_start_measurement(struct sophgo_saradc *saradc,
> + int channel)
> +{
> + writel(0, saradc->regs + SOPHGO_SARADC_CTRL_REG);
> + writel(SOPHGO_SARADC_SEL(channel) | SOPHGO_SARADC_EN,
> + saradc->regs + SOPHGO_SARADC_CTRL_REG);
> +}
> +
> +static int sophgo_saradc_wait(struct sophgo_saradc *saradc)
> +{
> + if (saradc->irq < 0) {
> + u32 reg;
> +
> + return readl_poll_timeout(saradc->regs + SOPHGO_SARADC_STATUS_REG,
> + reg, !(reg & SOPHGO_SARADC_BUSY),
> + 500, 1000000);
> + } else {
Returned above, so no need for else which helps with indent.
> + int ret;
> +
> + ret = wait_for_completion_timeout(&saradc->completion,
> + msecs_to_jiffies(1000)) > 0
> + ? 0 : -ETIMEDOUT;
> + return ret;
return wait_for_completion.
> + }
> +}
> +
> +static int sophgo_saradc_read_raw(struct iio_dev *indio_dev,
> + struct iio_chan_spec const *chan,
> + int *val, int *val2, long mask)
> +{
> + switch (mask) {
> + case IIO_CHAN_INFO_RAW:
> + struct sophgo_saradc *saradc = iio_priv(indio_dev);
> + u32 sample;
> + int ret;
> +
> + mutex_lock(&saradc->lock);
Use scoped_guard() here so you can return directly on error.
> + sophgo_saradc_start_measurement(saradc, chan->scan_index);
> + ret = sophgo_saradc_wait(saradc);
> + if (ret < 0) {
> + mutex_unlock(&saradc->lock);
> + return ret;
> + }
> +
> + sample = readl(saradc->regs + SOPHGO_SARADC_CH_RESULT_REG(chan->scan_index));
> + mutex_unlock(&saradc->lock);
> +
> + if (SARADC_CH_VALID(sample)) {
> + *val = SARADC_CH_RESULT(sample);
> + return IIO_VAL_INT;
> + }
errors out of line preferred.
if (!SARADC_CH_VALID(sample))
return -ENODATA;
*val...
> + return -ENODATA;
> + case IIO_CHAN_INFO_SCALE:
> + *val = 3300;
> + *val2 = 12;
> + return IIO_VAL_FRACTIONAL_LOG2;
> + default:
> + return -EINVAL;
> + }
> +}
> +
> +static irqreturn_t sophgo_saradc_interrupt_handler(int irq, void *dev_id)
> +{
> + struct sophgo_saradc *saradc = dev_id;
dev_id is an odd name for something that isn't an id.
No register to confirm the interupt status? Fine if not, but relatively
unusual.
> +
> + writel(1, saradc->regs + SOPHGO_SARADC_INTR_CLR_REG);
> + complete(&saradc->completion);
> + return IRQ_HANDLED;
> +}
> +
Single blank line
> +
> +static const struct of_device_id sophgo_saradc_match[] = {
> + { .compatible = "sophgo,cv18xx-saradc", },
> + { },
No comma needed on 'null' terminators like this as we won't
ever add anything after them.
> +};
> +MODULE_DEVICE_TABLE(of, sophgo_saradc_match);
> +
> +static int sophgo_saradc_probe(struct platform_device *pdev)
> +{
> + struct sophgo_saradc *saradc;
> + struct iio_dev *indio_dev;
> + int ret;
> +
> + indio_dev = devm_iio_device_alloc(&pdev->dev, sizeof(*saradc));
> + if (!indio_dev)
> + return -ENOMEM;
> +
> + saradc = iio_priv(indio_dev);
> + indio_dev->name = "Sophgo SARADC";
Hmm. Whilst the ABI docs don't say it, I don't think we have names with
either capitals or spaces. sophgo-saradc or similar preferred.
> + indio_dev->info = &saradc->info;
This is making some static const data dynamic. I can't see why you'd do this
> + indio_dev->modes = INDIO_DIRECT_MODE;
> + indio_dev->num_channels = 3;
ARRAY_SIZE(sophgo_channels)
> + indio_dev->channels = sophgo_channels;
> +
> + saradc->clk = devm_clk_get(&pdev->dev, NULL);
If it's optional, then use dev_clk_get_optional()
> + if (IS_ERR(saradc->clk)) {
> + dev_dbg(&pdev->dev, "Can't get clock from device tree, using No-Die domain");
> + } else {
> + ret = clk_prepare_enable(saradc->clk);
Perhaps devm_clk_get_optional_enabled() is appropriate, then
you can use null return to detect that it wasn't present - vs an ERR_PTR() for
it was but didn't enable.
> + if (ret)
> + return ret;
> + }
> +
> + saradc->regs = devm_platform_ioremap_resource(pdev, 0);
> + if (IS_ERR(saradc->regs)) {
> + ret = PTR_ERR(saradc->regs);
> + goto error_disable_clock;
> + }
> +
> + saradc->irq = platform_get_irq_optional(pdev, 0);
> + if (saradc->irq >= 0) {
> + ret = devm_request_irq(&pdev->dev, saradc->irq,
> + sophgo_saradc_interrupt_handler, 0,
> + dev_name(&pdev->dev), saradc);
> + if (ret)
> + goto error_disable_clock;
> +
> + writel(1, saradc->regs + SOPHGO_SARADC_INTR_EN_REG);
> +
> + init_completion(&saradc->completion);
Usually best to init a completion that is irq related before requesting
the irq. Avoids any possible races.
> + }
> +
> + saradc->info.read_raw = &sophgo_saradc_read_raw;
Why?
> +
> + mutex_init(&saradc->lock);
> + platform_set_drvdata(pdev, indio_dev);
> + writel(SOPHGO_SARADC_DEF_CYC_SETTINGS, saradc->regs + SOPHGO_SARADC_CYC_SET_REG);
> + ret = devm_iio_device_register(&pdev->dev, indio_dev);
> + if (ret)
> + goto error_disable_clock;
> +
> + return 0;
> +
> +error_disable_clock:
> + if (!IS_ERR(saradc->clk))
Don't mix and match devm and non devm management.
As above, I think you can just use devm_clk_get_optional_enabled()
> + clk_disable_unprepare(saradc->clk);
> + return ret;
> +}
> +
> +static void sophgo_saradc_remove(struct platform_device *pdev)
> +{
> + struct iio_dev *indio_dev = platform_get_drvdata(pdev);
> + struct sophgo_saradc *saradc = iio_priv(indio_dev);
> +
> + if (!IS_ERR(saradc->clk))
> + clk_disable_unprepare(saradc->clk);
> +}
> +
Single blank line is enough.
> +
> +static struct platform_driver sophgo_saradc_driver = {
> + .driver = {
> + .name = "sophgo-saradc",
> + .of_match_table = sophgo_saradc_match,
> + },
> + .probe = sophgo_saradc_probe,
> + .remove_new = sophgo_saradc_remove,
> +};
> +module_platform_driver(sophgo_saradc_driver);
> +
> +MODULE_AUTHOR("Thomas Bonnefille <thomas.bonnefille@...tlin.com>");
> +MODULE_DESCRIPTION("Sophgo SARADC driver");
> +MODULE_LICENSE("GPL");
>
Powered by blists - more mailing lists