All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stephen Boyd <sboyd@kernel.org>
To: Daniel Mack <daniel@zonque.org>, mturquette@baylibre.com
Cc: linux-clk@vger.kernel.org, robh@kernel.org,
	devicetree@vger.kernel.org, Daniel Mack <zonque@gmail.com>
Subject: Re: [PATCH v2 2/2] clk: Add driver for MAX9485
Date: Fri, 01 Jun 2018 23:13:58 -0700	[thread overview]
Message-ID: <152792003850.225090.13213694533140196613@swboyd.mtv.corp.google.com> (raw)
In-Reply-To: <20180525182058.25969-3-daniel@zonque.org>

Quoting Daniel Mack (2018-05-25 11:20:58)
> diff --git a/drivers/clk/Kconfig b/drivers/clk/Kconfig
> index 34968a381d0f..f8ab54c41d16 100644
> --- a/drivers/clk/Kconfig
> +++ b/drivers/clk/Kconfig
> @@ -45,6 +45,14 @@ config COMMON_CLK_MAX77686
>           This driver supports Maxim 77620/77686/77802 crystal oscillator
>           clock.
>  
> +config COMMON_CLK_MAX9485
> +       tristate "Clock driver for Maxim 9485 Programmable Clock Generator"
> +       depends on I2C
> +       depends on OF

Drop this depends? Probably not needed to compile this driver.

> +       depends on GPIOLIB || COMPILE_TEST

Is this common? I never knew.

> +       ---help---
> +         This driver supports Maxim 9485 Programmable Audio Clock Generator
> +
>  config COMMON_CLK_RK808
>         tristate "Clock driver for RK805/RK808/RK818"
>         depends on MFD_RK808
> diff --git a/drivers/clk/clk-max9485.c b/drivers/clk/clk-max9485.c
> new file mode 100644
> index 000000000000..9bbf2fee49c5
> --- /dev/null
> +++ b/drivers/clk/clk-max9485.c
> @@ -0,0 +1,408 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * clk-max9485.c: MAX9485 Programmable Audio Clock Generator
> + *
> + * (c) 2018 Daniel Mack <daniel@zonque.org>
> + *
> + * References:
> + *   MAX9485 Datasheet
> + *     http://www.maximintegrated.com/datasheet/index.mvp/id/4421
> + *
> + * This program is free software; you can redistribute  it and/or modify it
> + * under  the terms of  the GNU General  Public License as published by the
> + * Free Software Foundation;  either version 2 of the  License, or (at your
> + * option) any later version.
> + */
> +
> +#include <linux/module.h>
> +#include <linux/kernel.h>
> +#include <linux/clkdev.h>

Is this used?

> +#include <linux/clk.h>
> +#include <linux/clk-provider.h>
> +#include <linux/err.h>
> +#include <linux/errno.h>
> +#include <linux/gpio/consumer.h>
> +#include <linux/i2c.h>
> +#include <linux/regulator/consumer.h>
> +
> +#include <dt-bindings/clock/maxim,max9485.h>
> +
> +#define MAX9485_INPUT_FREQ 27000000UL
> +#define MAX9485_NUM_CLKS 4
> +
> +/* This chip has only one register of 8 bit width. */

So efficient!

> +
> +enum {
> +       MAX9485_FS_12KHZ        = 0 << 0,
> +       MAX9485_FS_32KHZ        = 1 << 0,
> +       MAX9485_FS_44_1KHZ      = 2 << 0,
> +       MAX9485_FS_48KHZ        = 3 << 0,
> +};
> +
> +enum {
> +       MAX9485_SCALE_256       = 0 << 2,
> +       MAX9485_SCALE_384       = 1 << 2,
> +       MAX9485_SCALE_768       = 2 << 2,
> +};

Any reason to be an enum? Maybe they can just be #defines.

> +
> +#define MAX9485_DOUBLE         BIT(4)
> +#define MAX9485_CLKOUT1_ENABLE BIT(5)
> +#define MAX9485_CLKOUT2_ENABLE BIT(6)
> +#define MAX9485_MCLK_ENABLE    BIT(7)
> +#define MAX9485_FREQ_MASK      0x1f
> +
> +struct max9485_rate {
> +       unsigned long out;
> +       u8 reg_value;
> +};
> +
> +/*
> + * Ordered by frequency. For frequency the hardware can generate with
> + * multiple settings, only the one with lowest jitter is listed.
> + */
> +static const struct max9485_rate max9485_rates[] = {
> +       {  3072000, MAX9485_FS_12KHZ   | MAX9485_SCALE_256 },
> +       {  4608000, MAX9485_FS_12KHZ   | MAX9485_SCALE_384 },
> +       {  8192000, MAX9485_FS_32KHZ   | MAX9485_SCALE_256 },
> +       {  9126000, MAX9485_FS_12KHZ   | MAX9485_SCALE_768 },
> +       { 11289600, MAX9485_FS_44_1KHZ | MAX9485_SCALE_256 },
> +       { 12288000, MAX9485_FS_48KHZ   | MAX9485_SCALE_256 },
> +       { 16384000, MAX9485_FS_32KHZ   | MAX9485_SCALE_256 | MAX9485_DOUBLE },
> +       { 16934400, MAX9485_FS_44_1KHZ | MAX9485_SCALE_384 },
> +       { 18384000, MAX9485_FS_48KHZ   | MAX9485_SCALE_384 },
> +       { 22579200, MAX9485_FS_44_1KHZ | MAX9485_SCALE_256 | MAX9485_DOUBLE },
> +       { 24576000, MAX9485_FS_48KHZ   | MAX9485_SCALE_256 | MAX9485_DOUBLE },
> +       { 33868800, MAX9485_FS_44_1KHZ | MAX9485_SCALE_384 | MAX9485_DOUBLE },
> +       { 36864000, MAX9485_FS_48KHZ   | MAX9485_SCALE_384 | MAX9485_DOUBLE },
> +       { 49152000, MAX9485_FS_32KHZ   | MAX9485_SCALE_768 | MAX9485_DOUBLE },
> +       { 67737600, MAX9485_FS_44_1KHZ | MAX9485_SCALE_768 | MAX9485_DOUBLE },
> +       { 73728000, MAX9485_FS_48KHZ   | MAX9485_SCALE_768 | MAX9485_DOUBLE },
> +       { } /* sentinel */
> +};
> +
> +struct max9485_driver_data;
> +
> +struct max9485_clk_hw {
> +       struct clk_hw hw;
> +       struct clk_init_data init;
> +       u8 enable_bit;
> +       struct max9485_driver_data *drvdata;
> +};
> +
> +struct max9485_driver_data {
> +       struct clk *xclk;
> +       struct i2c_client *client;
> +       u8 reg_value;
> +       unsigned long clkout_rate;
> +       struct regulator *supply;
> +       struct gpio_desc *reset_gpio;
> +       struct max9485_clk_hw hw[MAX9485_NUM_CLKS];
> +
> +       struct {
> +               struct clk_hw_onecell_data data;
> +               struct clk_hw *hw[MAX9485_NUM_CLKS];

I don't quite understand this one. There's already an array of clk_hw
pointers inside of clk_hw_onecell_data so just use that? But then
there's also an array of max9485_clk_hw structures, so maybe roll your
own clk_hw getter function?

> +       } onecell;
> +};
> +
> +static int max9485_update_bits(struct max9485_driver_data *drvdata,
> +                              u8 mask, u8 value)
> +{
> +       int ret;
> +
> +       drvdata->reg_value &= ~mask;
> +       drvdata->reg_value |= value;
> +
> +       dev_dbg(&drvdata->client->dev,
> +               "updating mask %02x value %02x -> %02x\n",
> +               mask, value, drvdata->reg_value);
> +
> +       ret = i2c_master_send(drvdata->client,
> +                             &drvdata->reg_value,
> +                             sizeof(drvdata->reg_value));

Maybe use a regmap instead? Then you could use regmap_update_bits() in
the places you need it.

> +
> +       return (ret < 0) ? ret : 0;
> +}
> +
> +static int max9485_clk_prepare(struct clk_hw *hw)
> +{
> +       struct max9485_clk_hw *clk_hw =
> +               container_of(hw, struct max9485_clk_hw, hw);

Please make a static inline function for this line called something like
to_max8458_clk().

> +
> +       return max9485_update_bits(clk_hw->drvdata,
> +                                  clk_hw->enable_bit,
> +                                  clk_hw->enable_bit);
> +}
> +
> +static void max9485_clk_unprepare(struct clk_hw *hw)
> +{
> +       struct max9485_clk_hw *clk_hw =
> +               container_of(hw, struct max9485_clk_hw, hw);
> +
> +       max9485_update_bits(clk_hw->drvdata, clk_hw->enable_bit, 0);
> +}
> +
> +/*
> + * MCLK OUT
> + */
> +
> +/* The MCLK output can only provide the same rate as the input clock */
> +static unsigned long max9485_mclkout_recalc_rate(struct clk_hw *hw,
> +                                                unsigned long parent_rate)
> +{
> +       return MAX9485_INPUT_FREQ;
> +}

This clk_op can be dropped and just rely on parent frequency.

> +
> +/*
> + * CLKOUT - configurable clock output
> + */
> +static int max9485_clkout_set_rate(struct clk_hw *hw, unsigned long rate,
> +                                  unsigned long parent_rate)
> +{
> +       const struct max9485_rate *entry;
> +       struct max9485_clk_hw *clk_hw =
> +               container_of(hw, struct max9485_clk_hw, hw);
> +
> +       for (entry = max9485_rates; entry->out != 0; entry++)
> +               if (entry->out == rate)
> +                       break;
> +
> +       if (entry->out == 0)
> +               return -EINVAL;
> +
> +       clk_hw->drvdata->clkout_rate = rate;
> +
> +       return max9485_update_bits(clk_hw->drvdata,
> +                                  MAX9485_FREQ_MASK,
> +                                  entry->reg_value);

Need to unwind clkout_rate on failure? clkout_rate should go away
actually.

> +}
> +
> +static long max9485_clkout_round_rate(struct clk_hw *hw, unsigned long rate,
> +                                     unsigned long *parent_rate)
> +{
> +       const struct max9485_rate *curr, *prev = NULL;
> +
> +       for (curr = max9485_rates; curr->out != 0; curr++) {
> +               /* Exact matches */
> +               if (curr->out == rate)
> +                       return rate;
> +
> +               /*
> +                * Find the first entry that has a frequency higher than the
> +                * requested one.
> +                */
> +               if (curr->out > rate) {
> +                       unsigned int mid;
> +
> +                       /*
> +                        * If this is the first entry, clamp the value to the
> +                        * lowest possible frequency.
> +                        */
> +                       if (!prev)
> +                               return curr->out;
> +
> +                       /*
> +                        * Otherwise, determine whether the previous entry or
> +                        * current one is closer.
> +                        */
> +                       mid = prev->out + ((curr->out - prev->out) / 2);
> +
> +                       return (mid > rate) ? prev->out : curr->out;
> +               }
> +
> +               prev = curr;
> +       }
> +
> +       /* If the last entry was still too high, clamp the value */
> +       return prev->out;
> +}
> +
> +static unsigned long max9485_clkout_recalc_rate(struct clk_hw *hw,
> +                                               unsigned long parent_rate)
> +{
> +       struct max9485_clk_hw *clk_hw =
> +               container_of(hw, struct max9485_clk_hw, hw);
> +
> +       return clk_hw->drvdata->clkout_rate;

This needs to return the rate of the clk that's determined by reading
the hardware. Caching the rate yourself is hard and error prone, so you
should only do it if you really have to.

> +}
> +
> +struct max9485_clk {
> +       const char *name;
> +       int parent_index;
> +       const struct clk_ops ops;
> +       u8 enable_bit;
> +};
> +
> +static const struct max9485_clk max9485_clks[MAX9485_NUM_CLKS] = {
> +       [MAX9485_MCLKOUT] = {
> +               .name = "mclkout",
> +               .parent_index = -1,
> +               .enable_bit = MAX9485_MCLK_ENABLE,
> +               .ops = {
> +                       .prepare        = max9485_clk_prepare,
> +                       .unprepare      = max9485_clk_unprepare,
> +                       .recalc_rate    = max9485_mclkout_recalc_rate,
> +               },
> +       },
> +       [MAX9485_CLKOUT] = {
> +               .name = "clkout",
> +               .parent_index = -1,
> +               .ops = {
> +                       .set_rate       = max9485_clkout_set_rate,
> +                       .round_rate     = max9485_clkout_round_rate,
> +                       .recalc_rate    = max9485_clkout_recalc_rate,
> +               },
> +       },
> +       [MAX9485_CLKOUT1] = {
> +               .name = "clkout1",
> +               .parent_index = MAX9485_CLKOUT,
> +               .enable_bit = MAX9485_CLKOUT1_ENABLE,
> +               .ops = {
> +                       .prepare        = max9485_clk_prepare,
> +                       .unprepare      = max9485_clk_unprepare,
> +                       .recalc_rate    = max9485_clkout_recalc_rate,
> +               },
> +       },
> +       [MAX9485_CLKOUT2] = {
> +               .name = "clkout2",
> +               .parent_index = MAX9485_CLKOUT,
> +               .enable_bit = MAX9485_CLKOUT2_ENABLE,
> +               .ops = {
> +                       .prepare        = max9485_clk_prepare,
> +                       .unprepare      = max9485_clk_unprepare,
> +                       .recalc_rate    = max9485_clkout_recalc_rate,
> +               },
> +       },
> +};
> +
> +static int max9485_i2c_probe(struct i2c_client *client,
> +                            const struct i2c_device_id *id)
> +{
> +       struct max9485_driver_data *drvdata;
> +       struct device *dev = &client->dev;
> +       unsigned long freq;
> +       int i, ret;
> +
> +       drvdata = devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL);
> +       if (drvdata == NULL)

More standard form is 

	if (!drvdata)
		return -ENOMEM;

> +               return -ENOMEM;
> +
> +       drvdata->xclk = devm_clk_get(dev, "xclk");
> +       if (IS_ERR(drvdata->xclk))
> +               return PTR_ERR(drvdata->xclk);
> +
> +       freq = clk_get_rate(drvdata->xclk);
> +       if (freq != MAX9485_INPUT_FREQ) {
> +               dev_err(dev, "Illegal xclk frequency of %ld\n", freq);
> +               return -EINVAL;
> +       }

Is this necessary? Seems pretty useless to do everywhere to make sure
that someone didn't design the hardware incorrectly. It would be better
to return the rate of the parent "xclk" by not having a recalc_rate hook
and relying on the parent's rate.

> +
> +       drvdata->supply = devm_regulator_get(dev, "vdd");
> +       if (IS_ERR(drvdata->supply))
> +               return PTR_ERR(drvdata->supply);
> +
> +       ret = regulator_enable(drvdata->supply);
> +       if (ret < 0)
> +               return ret;
> +
> +       drvdata->reset_gpio =
> +               devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_HIGH);
> +       if (IS_ERR(drvdata->reset_gpio))
> +               return PTR_ERR(drvdata->reset_gpio);
> +
> +       i2c_set_clientdata(client, drvdata);
> +       drvdata->client = client;
> +
> +       for (i = 0; i < MAX9485_NUM_CLKS; i++) {
> +               int parent_index = max9485_clks[i].parent_index;
> +               const char *name;
> +
> +               if (of_property_read_string_index(dev->of_node,
> +                                                 "clock-output-names",
> +                                                 i, &name) == 0)
> +                       drvdata->hw[i].init.name = name;
> +               else
> +                       drvdata->hw[i].init.name = max9485_clks[i].name;

Add braces on both arms please.

> +
> +               drvdata->hw[i].init.ops = &max9485_clks[i].ops;
> +               drvdata->hw[i].init.flags = CLK_IS_BASIC;

Don't use this flag unless you need it. Let me go add a comment to that
effect to the code right now.

> +
> +               if (parent_index > 0) {
> +                       drvdata->hw[i].init.parent_names =
> +                               &drvdata->hw[parent_index].init.name;
> +                       drvdata->hw[i].init.num_parents = 1;
> +                       drvdata->hw[i].init.flags |= CLK_SET_RATE_PARENT;
> +               }
> +
> +               drvdata->hw[i].enable_bit = max9485_clks[i].enable_bit;
> +               drvdata->hw[i].hw.init = &drvdata->hw[i].init;
> +               drvdata->hw[i].drvdata = drvdata;
> +
> +               ret = devm_clk_hw_register(dev, &drvdata->hw[i].hw);
> +               if (ret < 0)
> +                       return ret;
> +
> +               drvdata->onecell.hw[i] = &drvdata->hw[i].hw;
> +       }
> +
> +       drvdata->onecell.data.num = MAX9485_NUM_CLKS;
> +
> +       return devm_of_clk_add_hw_provider(dev, of_clk_hw_onecell_get,
> +                                          &drvdata->onecell.data);
> +}
> +
> +static int __maybe_unused max9485_suspend(struct device *dev)
> +{
> +       struct i2c_client *client = to_i2c_client(dev);
> +       struct max9485_driver_data *drvdata = i2c_get_clientdata(client);
> +
> +       if (drvdata->reset_gpio)
> +               gpiod_set_value_cansleep(drvdata->reset_gpio, 0);

Looks like you can call this unconditionally if the gpio is not there
because get_optional will return NULL and this function will bail out
early if so.

> +
> +       return 0;
> +}
> +
> +static int __maybe_unused max9485_resume(struct device *dev)
> +{
> +       struct i2c_client *client = to_i2c_client(dev);
> +       struct max9485_driver_data *drvdata = i2c_get_clientdata(client);
> +       int ret;
> +
> +       if (drvdata->reset_gpio)
> +               gpiod_set_value_cansleep(drvdata->reset_gpio, 0);
> +

Same.

> +       ret = i2c_master_send(client, &drvdata->reg_value,
> +                             sizeof(drvdata->reg_value));
> +
> +       return (ret < 0) ? ret : 0;

Drop the extra parenthesis please.

> +}
> +

WARNING: multiple messages have this Message-ID (diff)
From: Stephen Boyd <sboyd@kernel.org>
To: Daniel Mack <daniel@zonque.org>, mturquette@baylibre.com
Cc: linux-clk@vger.kernel.org, robh@kernel.org,
	devicetree@vger.kernel.org, Daniel Mack <zonque@gmail.com>,
	Daniel Mack <daniel@zonque.org>
Subject: Re: [PATCH v2 2/2] clk: Add driver for MAX9485
Date: Fri, 01 Jun 2018 23:13:58 -0700	[thread overview]
Message-ID: <152792003850.225090.13213694533140196613@swboyd.mtv.corp.google.com> (raw)
In-Reply-To: <20180525182058.25969-3-daniel@zonque.org>

Quoting Daniel Mack (2018-05-25 11:20:58)
> diff --git a/drivers/clk/Kconfig b/drivers/clk/Kconfig
> index 34968a381d0f..f8ab54c41d16 100644
> --- a/drivers/clk/Kconfig
> +++ b/drivers/clk/Kconfig
> @@ -45,6 +45,14 @@ config COMMON_CLK_MAX77686
>           This driver supports Maxim 77620/77686/77802 crystal oscillator
>           clock.
>  =

> +config COMMON_CLK_MAX9485
> +       tristate "Clock driver for Maxim 9485 Programmable Clock Generato=
r"
> +       depends on I2C
> +       depends on OF

Drop this depends? Probably not needed to compile this driver.

> +       depends on GPIOLIB || COMPILE_TEST

Is this common? I never knew.

> +       ---help---
> +         This driver supports Maxim 9485 Programmable Audio Clock Genera=
tor
> +
>  config COMMON_CLK_RK808
>         tristate "Clock driver for RK805/RK808/RK818"
>         depends on MFD_RK808
> diff --git a/drivers/clk/clk-max9485.c b/drivers/clk/clk-max9485.c
> new file mode 100644
> index 000000000000..9bbf2fee49c5
> --- /dev/null
> +++ b/drivers/clk/clk-max9485.c
> @@ -0,0 +1,408 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * clk-max9485.c: MAX9485 Programmable Audio Clock Generator
> + *
> + * (c) 2018 Daniel Mack <daniel@zonque.org>
> + *
> + * References:
> + *   MAX9485 Datasheet
> + *     http://www.maximintegrated.com/datasheet/index.mvp/id/4421
> + *
> + * This program is free software; you can redistribute  it and/or modify=
 it
> + * under  the terms of  the GNU General  Public License as published by =
the
> + * Free Software Foundation;  either version 2 of the  License, or (at y=
our
> + * option) any later version.
> + */
> +
> +#include <linux/module.h>
> +#include <linux/kernel.h>
> +#include <linux/clkdev.h>

Is this used?

> +#include <linux/clk.h>
> +#include <linux/clk-provider.h>
> +#include <linux/err.h>
> +#include <linux/errno.h>
> +#include <linux/gpio/consumer.h>
> +#include <linux/i2c.h>
> +#include <linux/regulator/consumer.h>
> +
> +#include <dt-bindings/clock/maxim,max9485.h>
> +
> +#define MAX9485_INPUT_FREQ 27000000UL
> +#define MAX9485_NUM_CLKS 4
> +
> +/* This chip has only one register of 8 bit width. */

So efficient!

> +
> +enum {
> +       MAX9485_FS_12KHZ        =3D 0 << 0,
> +       MAX9485_FS_32KHZ        =3D 1 << 0,
> +       MAX9485_FS_44_1KHZ      =3D 2 << 0,
> +       MAX9485_FS_48KHZ        =3D 3 << 0,
> +};
> +
> +enum {
> +       MAX9485_SCALE_256       =3D 0 << 2,
> +       MAX9485_SCALE_384       =3D 1 << 2,
> +       MAX9485_SCALE_768       =3D 2 << 2,
> +};

Any reason to be an enum? Maybe they can just be #defines.

> +
> +#define MAX9485_DOUBLE         BIT(4)
> +#define MAX9485_CLKOUT1_ENABLE BIT(5)
> +#define MAX9485_CLKOUT2_ENABLE BIT(6)
> +#define MAX9485_MCLK_ENABLE    BIT(7)
> +#define MAX9485_FREQ_MASK      0x1f
> +
> +struct max9485_rate {
> +       unsigned long out;
> +       u8 reg_value;
> +};
> +
> +/*
> + * Ordered by frequency. For frequency the hardware can generate with
> + * multiple settings, only the one with lowest jitter is listed.
> + */
> +static const struct max9485_rate max9485_rates[] =3D {
> +       {  3072000, MAX9485_FS_12KHZ   | MAX9485_SCALE_256 },
> +       {  4608000, MAX9485_FS_12KHZ   | MAX9485_SCALE_384 },
> +       {  8192000, MAX9485_FS_32KHZ   | MAX9485_SCALE_256 },
> +       {  9126000, MAX9485_FS_12KHZ   | MAX9485_SCALE_768 },
> +       { 11289600, MAX9485_FS_44_1KHZ | MAX9485_SCALE_256 },
> +       { 12288000, MAX9485_FS_48KHZ   | MAX9485_SCALE_256 },
> +       { 16384000, MAX9485_FS_32KHZ   | MAX9485_SCALE_256 | MAX9485_DOUB=
LE },
> +       { 16934400, MAX9485_FS_44_1KHZ | MAX9485_SCALE_384 },
> +       { 18384000, MAX9485_FS_48KHZ   | MAX9485_SCALE_384 },
> +       { 22579200, MAX9485_FS_44_1KHZ | MAX9485_SCALE_256 | MAX9485_DOUB=
LE },
> +       { 24576000, MAX9485_FS_48KHZ   | MAX9485_SCALE_256 | MAX9485_DOUB=
LE },
> +       { 33868800, MAX9485_FS_44_1KHZ | MAX9485_SCALE_384 | MAX9485_DOUB=
LE },
> +       { 36864000, MAX9485_FS_48KHZ   | MAX9485_SCALE_384 | MAX9485_DOUB=
LE },
> +       { 49152000, MAX9485_FS_32KHZ   | MAX9485_SCALE_768 | MAX9485_DOUB=
LE },
> +       { 67737600, MAX9485_FS_44_1KHZ | MAX9485_SCALE_768 | MAX9485_DOUB=
LE },
> +       { 73728000, MAX9485_FS_48KHZ   | MAX9485_SCALE_768 | MAX9485_DOUB=
LE },
> +       { } /* sentinel */
> +};
> +
> +struct max9485_driver_data;
> +
> +struct max9485_clk_hw {
> +       struct clk_hw hw;
> +       struct clk_init_data init;
> +       u8 enable_bit;
> +       struct max9485_driver_data *drvdata;
> +};
> +
> +struct max9485_driver_data {
> +       struct clk *xclk;
> +       struct i2c_client *client;
> +       u8 reg_value;
> +       unsigned long clkout_rate;
> +       struct regulator *supply;
> +       struct gpio_desc *reset_gpio;
> +       struct max9485_clk_hw hw[MAX9485_NUM_CLKS];
> +
> +       struct {
> +               struct clk_hw_onecell_data data;
> +               struct clk_hw *hw[MAX9485_NUM_CLKS];

I don't quite understand this one. There's already an array of clk_hw
pointers inside of clk_hw_onecell_data so just use that? But then
there's also an array of max9485_clk_hw structures, so maybe roll your
own clk_hw getter function?

> +       } onecell;
> +};
> +
> +static int max9485_update_bits(struct max9485_driver_data *drvdata,
> +                              u8 mask, u8 value)
> +{
> +       int ret;
> +
> +       drvdata->reg_value &=3D ~mask;
> +       drvdata->reg_value |=3D value;
> +
> +       dev_dbg(&drvdata->client->dev,
> +               "updating mask %02x value %02x -> %02x\n",
> +               mask, value, drvdata->reg_value);
> +
> +       ret =3D i2c_master_send(drvdata->client,
> +                             &drvdata->reg_value,
> +                             sizeof(drvdata->reg_value));

Maybe use a regmap instead? Then you could use regmap_update_bits() in
the places you need it.

> +
> +       return (ret < 0) ? ret : 0;
> +}
> +
> +static int max9485_clk_prepare(struct clk_hw *hw)
> +{
> +       struct max9485_clk_hw *clk_hw =3D
> +               container_of(hw, struct max9485_clk_hw, hw);

Please make a static inline function for this line called something like
to_max8458_clk().

> +
> +       return max9485_update_bits(clk_hw->drvdata,
> +                                  clk_hw->enable_bit,
> +                                  clk_hw->enable_bit);
> +}
> +
> +static void max9485_clk_unprepare(struct clk_hw *hw)
> +{
> +       struct max9485_clk_hw *clk_hw =3D
> +               container_of(hw, struct max9485_clk_hw, hw);
> +
> +       max9485_update_bits(clk_hw->drvdata, clk_hw->enable_bit, 0);
> +}
> +
> +/*
> + * MCLK OUT
> + */
> +
> +/* The MCLK output can only provide the same rate as the input clock */
> +static unsigned long max9485_mclkout_recalc_rate(struct clk_hw *hw,
> +                                                unsigned long parent_rat=
e)
> +{
> +       return MAX9485_INPUT_FREQ;
> +}

This clk_op can be dropped and just rely on parent frequency.

> +
> +/*
> + * CLKOUT - configurable clock output
> + */
> +static int max9485_clkout_set_rate(struct clk_hw *hw, unsigned long rate,
> +                                  unsigned long parent_rate)
> +{
> +       const struct max9485_rate *entry;
> +       struct max9485_clk_hw *clk_hw =3D
> +               container_of(hw, struct max9485_clk_hw, hw);
> +
> +       for (entry =3D max9485_rates; entry->out !=3D 0; entry++)
> +               if (entry->out =3D=3D rate)
> +                       break;
> +
> +       if (entry->out =3D=3D 0)
> +               return -EINVAL;
> +
> +       clk_hw->drvdata->clkout_rate =3D rate;
> +
> +       return max9485_update_bits(clk_hw->drvdata,
> +                                  MAX9485_FREQ_MASK,
> +                                  entry->reg_value);

Need to unwind clkout_rate on failure? clkout_rate should go away
actually.

> +}
> +
> +static long max9485_clkout_round_rate(struct clk_hw *hw, unsigned long r=
ate,
> +                                     unsigned long *parent_rate)
> +{
> +       const struct max9485_rate *curr, *prev =3D NULL;
> +
> +       for (curr =3D max9485_rates; curr->out !=3D 0; curr++) {
> +               /* Exact matches */
> +               if (curr->out =3D=3D rate)
> +                       return rate;
> +
> +               /*
> +                * Find the first entry that has a frequency higher than =
the
> +                * requested one.
> +                */
> +               if (curr->out > rate) {
> +                       unsigned int mid;
> +
> +                       /*
> +                        * If this is the first entry, clamp the value to=
 the
> +                        * lowest possible frequency.
> +                        */
> +                       if (!prev)
> +                               return curr->out;
> +
> +                       /*
> +                        * Otherwise, determine whether the previous entr=
y or
> +                        * current one is closer.
> +                        */
> +                       mid =3D prev->out + ((curr->out - prev->out) / 2);
> +
> +                       return (mid > rate) ? prev->out : curr->out;
> +               }
> +
> +               prev =3D curr;
> +       }
> +
> +       /* If the last entry was still too high, clamp the value */
> +       return prev->out;
> +}
> +
> +static unsigned long max9485_clkout_recalc_rate(struct clk_hw *hw,
> +                                               unsigned long parent_rate)
> +{
> +       struct max9485_clk_hw *clk_hw =3D
> +               container_of(hw, struct max9485_clk_hw, hw);
> +
> +       return clk_hw->drvdata->clkout_rate;

This needs to return the rate of the clk that's determined by reading
the hardware. Caching the rate yourself is hard and error prone, so you
should only do it if you really have to.

> +}
> +
> +struct max9485_clk {
> +       const char *name;
> +       int parent_index;
> +       const struct clk_ops ops;
> +       u8 enable_bit;
> +};
> +
> +static const struct max9485_clk max9485_clks[MAX9485_NUM_CLKS] =3D {
> +       [MAX9485_MCLKOUT] =3D {
> +               .name =3D "mclkout",
> +               .parent_index =3D -1,
> +               .enable_bit =3D MAX9485_MCLK_ENABLE,
> +               .ops =3D {
> +                       .prepare        =3D max9485_clk_prepare,
> +                       .unprepare      =3D max9485_clk_unprepare,
> +                       .recalc_rate    =3D max9485_mclkout_recalc_rate,
> +               },
> +       },
> +       [MAX9485_CLKOUT] =3D {
> +               .name =3D "clkout",
> +               .parent_index =3D -1,
> +               .ops =3D {
> +                       .set_rate       =3D max9485_clkout_set_rate,
> +                       .round_rate     =3D max9485_clkout_round_rate,
> +                       .recalc_rate    =3D max9485_clkout_recalc_rate,
> +               },
> +       },
> +       [MAX9485_CLKOUT1] =3D {
> +               .name =3D "clkout1",
> +               .parent_index =3D MAX9485_CLKOUT,
> +               .enable_bit =3D MAX9485_CLKOUT1_ENABLE,
> +               .ops =3D {
> +                       .prepare        =3D max9485_clk_prepare,
> +                       .unprepare      =3D max9485_clk_unprepare,
> +                       .recalc_rate    =3D max9485_clkout_recalc_rate,
> +               },
> +       },
> +       [MAX9485_CLKOUT2] =3D {
> +               .name =3D "clkout2",
> +               .parent_index =3D MAX9485_CLKOUT,
> +               .enable_bit =3D MAX9485_CLKOUT2_ENABLE,
> +               .ops =3D {
> +                       .prepare        =3D max9485_clk_prepare,
> +                       .unprepare      =3D max9485_clk_unprepare,
> +                       .recalc_rate    =3D max9485_clkout_recalc_rate,
> +               },
> +       },
> +};
> +
> +static int max9485_i2c_probe(struct i2c_client *client,
> +                            const struct i2c_device_id *id)
> +{
> +       struct max9485_driver_data *drvdata;
> +       struct device *dev =3D &client->dev;
> +       unsigned long freq;
> +       int i, ret;
> +
> +       drvdata =3D devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL);
> +       if (drvdata =3D=3D NULL)

More standard form is =


	if (!drvdata)
		return -ENOMEM;

> +               return -ENOMEM;
> +
> +       drvdata->xclk =3D devm_clk_get(dev, "xclk");
> +       if (IS_ERR(drvdata->xclk))
> +               return PTR_ERR(drvdata->xclk);
> +
> +       freq =3D clk_get_rate(drvdata->xclk);
> +       if (freq !=3D MAX9485_INPUT_FREQ) {
> +               dev_err(dev, "Illegal xclk frequency of %ld\n", freq);
> +               return -EINVAL;
> +       }

Is this necessary? Seems pretty useless to do everywhere to make sure
that someone didn't design the hardware incorrectly. It would be better
to return the rate of the parent "xclk" by not having a recalc_rate hook
and relying on the parent's rate.

> +
> +       drvdata->supply =3D devm_regulator_get(dev, "vdd");
> +       if (IS_ERR(drvdata->supply))
> +               return PTR_ERR(drvdata->supply);
> +
> +       ret =3D regulator_enable(drvdata->supply);
> +       if (ret < 0)
> +               return ret;
> +
> +       drvdata->reset_gpio =3D
> +               devm_gpiod_get_optional(dev, "reset", GPIOD_OUT_HIGH);
> +       if (IS_ERR(drvdata->reset_gpio))
> +               return PTR_ERR(drvdata->reset_gpio);
> +
> +       i2c_set_clientdata(client, drvdata);
> +       drvdata->client =3D client;
> +
> +       for (i =3D 0; i < MAX9485_NUM_CLKS; i++) {
> +               int parent_index =3D max9485_clks[i].parent_index;
> +               const char *name;
> +
> +               if (of_property_read_string_index(dev->of_node,
> +                                                 "clock-output-names",
> +                                                 i, &name) =3D=3D 0)
> +                       drvdata->hw[i].init.name =3D name;
> +               else
> +                       drvdata->hw[i].init.name =3D max9485_clks[i].name;

Add braces on both arms please.

> +
> +               drvdata->hw[i].init.ops =3D &max9485_clks[i].ops;
> +               drvdata->hw[i].init.flags =3D CLK_IS_BASIC;

Don't use this flag unless you need it. Let me go add a comment to that
effect to the code right now.

> +
> +               if (parent_index > 0) {
> +                       drvdata->hw[i].init.parent_names =3D
> +                               &drvdata->hw[parent_index].init.name;
> +                       drvdata->hw[i].init.num_parents =3D 1;
> +                       drvdata->hw[i].init.flags |=3D CLK_SET_RATE_PAREN=
T;
> +               }
> +
> +               drvdata->hw[i].enable_bit =3D max9485_clks[i].enable_bit;
> +               drvdata->hw[i].hw.init =3D &drvdata->hw[i].init;
> +               drvdata->hw[i].drvdata =3D drvdata;
> +
> +               ret =3D devm_clk_hw_register(dev, &drvdata->hw[i].hw);
> +               if (ret < 0)
> +                       return ret;
> +
> +               drvdata->onecell.hw[i] =3D &drvdata->hw[i].hw;
> +       }
> +
> +       drvdata->onecell.data.num =3D MAX9485_NUM_CLKS;
> +
> +       return devm_of_clk_add_hw_provider(dev, of_clk_hw_onecell_get,
> +                                          &drvdata->onecell.data);
> +}
> +
> +static int __maybe_unused max9485_suspend(struct device *dev)
> +{
> +       struct i2c_client *client =3D to_i2c_client(dev);
> +       struct max9485_driver_data *drvdata =3D i2c_get_clientdata(client=
);
> +
> +       if (drvdata->reset_gpio)
> +               gpiod_set_value_cansleep(drvdata->reset_gpio, 0);

Looks like you can call this unconditionally if the gpio is not there
because get_optional will return NULL and this function will bail out
early if so.

> +
> +       return 0;
> +}
> +
> +static int __maybe_unused max9485_resume(struct device *dev)
> +{
> +       struct i2c_client *client =3D to_i2c_client(dev);
> +       struct max9485_driver_data *drvdata =3D i2c_get_clientdata(client=
);
> +       int ret;
> +
> +       if (drvdata->reset_gpio)
> +               gpiod_set_value_cansleep(drvdata->reset_gpio, 0);
> +

Same.

> +       ret =3D i2c_master_send(client, &drvdata->reg_value,
> +                             sizeof(drvdata->reg_value));
> +
> +       return (ret < 0) ? ret : 0;

Drop the extra parenthesis please.

> +}
> +

  parent reply	other threads:[~2018-06-02  6:13 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-25 18:20 [PATCH v2 0/2] clk: Add driver for MAX9485 Daniel Mack
2018-05-25 18:20 ` [PATCH v2 1/2] dts: clk: add devicetree bindings " Daniel Mack
2018-05-31  3:37   ` Rob Herring
2018-05-25 18:20 ` [PATCH v2 2/2] clk: Add driver " Daniel Mack
2018-05-31  5:28   ` Daniel Mack
2018-06-02  6:13   ` Stephen Boyd [this message]
2018-06-02  6:13     ` Stephen Boyd
2018-06-02 11:14     ` Daniel Mack

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=152792003850.225090.13213694533140196613@swboyd.mtv.corp.google.com \
    --to=sboyd@kernel.org \
    --cc=daniel@zonque.org \
    --cc=devicetree@vger.kernel.org \
    --cc=linux-clk@vger.kernel.org \
    --cc=mturquette@baylibre.com \
    --cc=robh@kernel.org \
    --cc=zonque@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.