Re: [RFC PATCH v1 13/13] watchdog: bd70528: Initial support for ROHM BD70528 watchdog block

[Guenter Roeck <linux@roeck-us.net>]

On Tue, Jan 22, 2019 at 07:10:23PM +0200, Matti Vaittinen wrote:
> On Tue, Jan 22, 2019 at 07:47:50AM -0800, Guenter Roeck wrote:
> > On Tue, Jan 22, 2019 at 11:48:36AM +0200, Matti Vaittinen wrote:
> > > Initial support for watchdog block included in ROHM BD70528
> > > power management IC.
> > > 
> > > Configurations for low power states are still to be checked.
> > > 
> > > Signed-off-by: Matti Vaittinen <matti.vaittinen@fi.rohmeurope.com>
> > > ---
> > >  drivers/watchdog/Kconfig       |  12 +++
> > >  drivers/watchdog/Makefile      |   1 +
> > >  drivers/watchdog/bd70528_wdt.c | 161 +++++++++++++++++++++++++++++++++++++++++
> > >  3 files changed, 174 insertions(+)
> > >  create mode 100644 drivers/watchdog/bd70528_wdt.c
> > > 
> > > diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
> > > index 57f017d74a97..f30e3a3e886e 100644
> > > --- a/drivers/watchdog/Kconfig
> > > +++ b/drivers/watchdog/Kconfig
> > > @@ -90,6 +90,18 @@ config SOFT_WATCHDOG_PRETIMEOUT
> > >  	  watchdog. Be aware that governors might affect the watchdog because it
> > >  	  is purely software, e.g. the panic governor will stall it!
> > >  
> > > +config BD70528_WATCHDOG
> > > +	tristate "ROHM BD70528 PMIC Watchdog"
> > > +	depends on MFD_ROHM_BD70528
> > > +	select WATCHDOG_CORE
> > > +	help
> > > +	  Support for the watchdog in the ROHM BD70528 PMIC. Watchdog trigger
> > > +	  cause system reset.
> > > +
> > > +	  Say Y here to include support for the ROHM BD70528 watchdog.
> > > +	  Alternatively say M to compile the driver as a module,
> > > +	  which will be called bd70528_wdt.
> > > +
> > >  config DA9052_WATCHDOG
> > >  	tristate "Dialog DA9052 Watchdog"
> > >  	depends on PMIC_DA9052 || COMPILE_TEST
> > > diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
> > > index a0917ef28e07..1ce87a3b1172 100644
> > > --- a/drivers/watchdog/Makefile
> > > +++ b/drivers/watchdog/Makefile
> > > @@ -204,6 +204,7 @@ obj-$(CONFIG_WATCHDOG_SUN4V)		+= sun4v_wdt.o
> > >  obj-$(CONFIG_XEN_WDT) += xen_wdt.o
> > >  
> > >  # Architecture Independent
> > > +obj-$(CONFIG_BD70528_WATCHDOG) += bd70528_wdt.o
> > >  obj-$(CONFIG_DA9052_WATCHDOG) += da9052_wdt.o
> > >  obj-$(CONFIG_DA9055_WATCHDOG) += da9055_wdt.o
> > >  obj-$(CONFIG_DA9062_WATCHDOG) += da9062_wdt.o
> > > diff --git a/drivers/watchdog/bd70528_wdt.c b/drivers/watchdog/bd70528_wdt.c
> > > new file mode 100644
> > > index 000000000000..e9a32566f595
> > > --- /dev/null
> > > +++ b/drivers/watchdog/bd70528_wdt.c
> > > @@ -0,0 +1,161 @@
> > > +// SPDX-License-Identifier: GPL-2.0
> > > +// Copyright (C) 2018 ROHM Semiconductors
> > > +// ROHM BD70528MWV watchdog driver
> > > +
> > > +#include <linux/bcd.h>
> > > +#include <linux/kernel.h>
> > > +#include <linux/mfd/rohm-bd70528.h>
> > > +#include <linux/module.h>
> > > +#include <linux/of.h>
> > > +#include <linux/platform_device.h>
> > > +#include <linux/regmap.h>
> > > +#include <linux/watchdog.h>
> > > +
> > > +static int bd70528_wdt_set(struct bd70528 *bd70528, int enable)
> > > +{
> > > +	int ret;
> > > +
> > > +	if (bd70528->rtc_timer_lock)
> > > +		mutex_lock(bd70528->rtc_timer_lock);
> > 
> > This looks awkward. I don't think the if() is necessary.
> 
> Right. Now when only bd70528 MFD driver uses this WDT this if is not
> required.
> 
That doesn't warrant the conditional. It just bloats the code.
If there is only one user, the mutex will always be acquired.

> > > +
> > > +	ret = bd70528->wdt_set(bd70528, enable, NULL);
> > > +
> > > +	if (bd70528->rtc_timer_lock)
> > > +		mutex_unlock(bd70528->rtc_timer_lock);
> > > +	return ret;
> > > +}
> > > +
> > > +static int bd70528_wdt_start(struct watchdog_device *wdt)
> > > +{
> > > +	struct bd70528 *bd70528 = watchdog_get_drvdata(wdt);
> > > +
> > > +	return bd70528_wdt_set(bd70528, 1);
> > > +}
> > > +
> > > +static int bd70528_wdt_stop(struct watchdog_device *wdt)
> > > +{
> > > +	struct bd70528 *bd70528 = watchdog_get_drvdata(wdt);
> > > +
> > > +	return bd70528_wdt_set(bd70528, 0);
> > > +}
> > > +
> > > +static int bd70528_wdt_set_timeout(struct watchdog_device *wdt,
> > > +				    unsigned int timeout)
> > > +{
> > > +	unsigned int hours;
> > > +	unsigned int minutes;
> > > +	unsigned int seconds;
> > > +	int ret;
> > > +	struct bd70528 *bd70528 = watchdog_get_drvdata(wdt);
> > > +
> > > +	seconds = timeout;
> > > +	hours = timeout / (60 * 60);
> > > +	/* Maximum timeout is 1h 59m 59s => hours is 1 or 0 */
> > > +	if (hours)
> > > +		seconds -= (60 * 60);
> > > +	minutes = seconds / 60;
> > > +	seconds = seconds % 60;
> > > +
> > > +	if (bd70528->rtc_timer_lock)
> > > +		mutex_lock(bd70528->rtc_timer_lock);
> > > +
> > > +	ret = bd70528->wdt_set(bd70528, 0, NULL);
> > > +	if (ret)
> > > +		goto out_unlock;
> > > +
> > > +	ret = regmap_update_bits(bd70528->chip.regmap, BD70528_REG_WDT_HOUR,
> > > +				 BD70528_MASK_WDT_HOUR, hours);
> > > +	if (ret) {
> > > +		dev_err(bd70528->chip.dev, "Failed to set WDT hours\n");
> > > +		goto out_en_unlock;
> > > +	}
> > > +	ret = regmap_update_bits(bd70528->chip.regmap, BD70528_REG_WDT_MINUTE,
> > > +				 BD70528_MASK_WDT_MINUTE, bin2bcd(minutes));
> > > +	if (ret) {
> > > +		dev_err(bd70528->chip.dev, "Failed to set WDT minutes\n");
> > > +		goto out_en_unlock;
> > > +	}
> > > +	ret = regmap_update_bits(bd70528->chip.regmap, BD70528_REG_WDT_SEC,
> > > +				 BD70528_MASK_WDT_SEC, bin2bcd(seconds));
> > > +	if (ret) {
> > > +		dev_err(bd70528->chip.dev, "Failed to set WDT seconds\n");
> > > +		goto out_en_unlock;
> > 
> > Unnecessary goto.
> 
> True. I'll drop this.
> 
> > 
> > > +	}
> > > +
> > > +out_en_unlock:
> > > +	ret = bd70528->wdt_set(bd70528, 1, NULL);
> > > +out_unlock:
> > > +	if (bd70528->rtc_timer_lock)
> > > +		mutex_lock(bd70528->rtc_timer_lock);
> > 
> > I don't think this code was ever tested.
> 
> Yep. This should be unlock. What comes to testingI'll quote the
> cover-sheet for the patch set:
> 
> > Currently only MFD core, clk, RTC and regulator portions are
> > somehow tested. The RFC series also include initial gpio, power-supply
> > and watchdog patches in order to provide better overview on chip
> > and to collect initial feedback. Reset and ADC are not supported by
> > this series.
> 
> I think having the wdt_set and rtc_timer_lock in MFD would have been
> completely mysterious if watchdog draft was not included =)
> 
> > > +static const struct watchdog_info bd70528_wdt_info = {
> > > +	.identity = "bd70528-wdt",
> > > +	.options = WDIOF_SETTIMEOUT | WDIOF_KEEPALIVEPING | WDIOF_MAGICCLOSE,
> > > +};
> > > +
> > > +static const struct watchdog_ops bd70528_wdt_ops = {
> > > +	.start		= bd70528_wdt_start,
> > > +	.stop		= bd70528_wdt_stop,
> > > +	/*
> > > +	 *  bd70528 WDT ping is same as enable. Eg, writing 'enable' to enabled
> > > +	 * WDT will restart the timeout
> > > +	 */
> > Unnecessary comment.
> > 
> 
> Ok. I will remove the comment if this is obvious to others. For me it
> was not obvious. I was first writing a separate ping and start functions
> untill I realized that it is the same operation. But this was my first
> WDT driver so I don't know if this is a normal for all WDTs.
> 
It is documented as part of the API.

"Most hardware that does not support this as a separate function uses the
start function to restart the watchdog timer hardware"

Repeating the API for each driver doesn't really add value.

> > > +	.ping		= bd70528_wdt_start,
> > > +	.set_timeout	= bd70528_wdt_set_timeout,
> > > +};
> > > +
> > > +/* Max time we can set is 1 hour, 59 minutes and 59 seconds */
> > > +#define WDT_MAX_MS ((2 * 60 * 60 - 1) * 1000)
> > > +/* Minimum time is 1 second */
> > > +#define WDT_MIN_MS 1000
> > > +static struct watchdog_device bd70528_wd = {
> > > +	.info = &bd70528_wdt_info,
> > > +	.ops =  &bd70528_wdt_ops,
> > > +	.min_hw_heartbeat_ms = WDT_MIN_MS,
> > > +	.max_hw_heartbeat_ms = WDT_MAX_MS,
> > > +};
> > > +
> > > +static int bd70528_wdt_probe(struct platform_device *pdev)
> > > +{
> > > +	struct bd70528 *tmp;
> > > +	struct bd70528 *bd70528;
> > > +	int ret;
> > > +
> > > +	tmp = dev_get_drvdata(pdev->dev.parent);
> > > +	if (!tmp) {
> > > +		dev_err(&pdev->dev, "No MFD driver data\n");
> > > +		return -EINVAL;
> > > +	}
> > > +	bd70528 = devm_kzalloc(&pdev->dev, sizeof(*bd70528), GFP_KERNEL);
> > > +	if (!bd70528)
> > > +		return -ENOMEM;
> > > +
> > > +	*bd70528 = *tmp;
> > > +	bd70528->chip.dev = &pdev->dev;
> > 
> > This is wrong.
> > You should not copy the parent's driver data but have local driver
> > data as needed which then points to the parent's driver data if
> > needed. I assume this is why the mutex is a pointer, but that
> > just shows that the whole approach is wrong.
> 
> Mutex is a pointer because we want to use same mutex from WDT and RTC.
> We can sure point to parent data but then we still need our own dev
> pointer. So we can have a struct with pointer to parent data and dev
> pointer - but I'm not at all sure it is any clearer.

As I said, that is wrong. To say it in plaintext, I won't accept
the driver if it copies the parent's driver data. The driver should
have and use its own driver data, and only maintain a pointer to
its parent's driver data. And most definitely you don't want to
copy and use any device data structure from the parent.

> > 
> > > +
> > > +	/*
> > > +	 * TODO: Set the initial state and timeout.
> > 
> > Confused. Why don't you just do it ?
> 
> I will. But it's not ready yet. I still wanted to include the WDT for
> this RFC. And I hope I could get the MFD core part included in Lee's
> tree at early phase so that the include/linux/mfd/rohm-bd70528.h would
> be in other sub trees when they're finalized for upstreaming.
> 
> > 
> > > +	 * See whether the low power states require special handling
> > > +	 */
> > > +	watchdog_set_drvdata(&bd70528_wd, bd70528);
> > 
> > At least in theory there can be more than one of those devices
> > in the system, since it is an i2c device. With that in mind, bd70528_wd
> > should be locally allocated.
> 
> Point taken, thanks.
> 
> > Also, bd70528_wd should be fully initialized. For example, the parent
> > device is not set.
> 
> Thanks for this point too =) I will see what are all the missing
> initializations before sending out the final version.
> 
> I do really appreciate that you see the trouble of doing the review and
> giving me the push to right direction!
> 
> Br,
> 	Matti Vaittinen
> 
> -- 
> Matti Vaittinen
> ROHM Semiconductors
> 
> ~~~ "I don't think so," said Rene Descartes.  Just then, he vanished ~~~