All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ariel D'Alessandro <ariel@vanguardiasur.com.ar>
To: Guenter Roeck <linux@roeck-us.net>
Cc: linux-watchdog@vger.kernel.org, wim@iguana.be,
	ezequiel@vanguardiasur.com.ar
Subject: Re: [PATCH 1/2] watchdog: NXP LPC18XX Windowed Watchdog Timer Driver
Date: Sun, 28 Jun 2015 15:13:23 -0300	[thread overview]
Message-ID: <55903943.10309@vanguardiasur.com.ar> (raw)
In-Reply-To: <20150626190700.GC23078@roeck-us.net>

Hi Guenter,
Thanks for your quick reply.

El 26/06/15 a las 16:07, Guenter Roeck escibió:
> Hi Ariel,
> 
> On Fri, Jun 26, 2015 at 03:24:31PM -0300, Ariel D'Alessandro wrote:
>> This commit adds support for the watchdog timer found in NXP LPC SoCs
>> family, which includes LPC18xx/LPC43xx. Other SoCs in that family may
>> share the same watchdog hardware.
>>
>> Watchdog driver registers a restart handler that will restart the system
>> by performing an incorrect feed after ensuring the watchdog is enabled in
>> reset mode.
>>
>> As watchdog cannot be disabled in hardware, driver's stop routine will
>> regularly send a keepalive ping using a timer.
>>
>> Signed-off-by: Ariel D'Alessandro <ariel@vanguardiasur.com.ar>
>> ---
>>  drivers/watchdog/Kconfig       |  11 ++
>>  drivers/watchdog/Makefile      |   1 +
>>  drivers/watchdog/lpc18xx_wdt.c | 340 +++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 352 insertions(+)
>>  create mode 100644 drivers/watchdog/lpc18xx_wdt.c
>>
>> diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
>> index 742fbbc..31100c2 100644
>> --- a/drivers/watchdog/Kconfig
>> +++ b/drivers/watchdog/Kconfig
>> @@ -547,6 +547,17 @@ config DIGICOLOR_WATCHDOG
>>  	  To compile this driver as a module, choose M here: the
>>  	  module will be called digicolor_wdt.
>>  
>> +config LPC18XX_WATCHDOG
>> +	tristate "LCP18XX Watchdog"
>> +	depends on ARCH_LPC18XX
>> +	select WATCHDOG_CORE
>> +	help
>> +	  Say Y here if to include support for the watchdog timer
>> +	  in NXP LPC SoCs family, which includes LPC18xx/LPC43xx
>> +	  processors.
>> +	  To compile this driver as a module, choose M here: the
>> +	  module will be called lpc18xx_wdt.
>> +
>>  # AVR32 Architecture
>>  
>>  config AT32AP700X_WDT
>> diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
>> index 59ea9a1..1b0ef48 100644
>> --- a/drivers/watchdog/Makefile
>> +++ b/drivers/watchdog/Makefile
>> @@ -66,6 +66,7 @@ obj-$(CONFIG_TEGRA_WATCHDOG) += tegra_wdt.o
>>  obj-$(CONFIG_MESON_WATCHDOG) += meson_wdt.o
>>  obj-$(CONFIG_MEDIATEK_WATCHDOG) += mtk_wdt.o
>>  obj-$(CONFIG_DIGICOLOR_WATCHDOG) += digicolor_wdt.o
>> +obj-$(CONFIG_LPC18XX_WATCHDOG) += lpc18xx_wdt.o
>>  
>>  # AVR32 Architecture
>>  obj-$(CONFIG_AT32AP700X_WDT) += at32ap700x_wdt.o
>> diff --git a/drivers/watchdog/lpc18xx_wdt.c b/drivers/watchdog/lpc18xx_wdt.c
>> new file mode 100644
>> index 0000000..80d68eb
>> --- /dev/null
>> +++ b/drivers/watchdog/lpc18xx_wdt.c
>> @@ -0,0 +1,340 @@
>> +/*
>> + * NXP LPC18xx Windowed Watchdog Timer (WWDT)
>> + *
> 
> What does "Windowed" stand for ? That term doesn't really mean anything
> to me. How does it differ from a non-windowed watchdog driver ?

"NXP LPC18xx Windowed Watchdog Timer" is the full specific name of the
Watchdog, as it's written in the LPC18xx User Manual (UM10430).

"Windowed" means that, besides the maximum, a minimum time-out period
can be set, requiring reload operation (feed) to occur between these two
limits.

Linux watchdog framework doesn't take this property into account, so
"Windowed" term may not have much significance here. Do you think it
should be removed?

> 
>> + * Copyright (c) 2015 Ariel D'Alessandro <ariel@vanguardiasur.com>
>> + *
>> + * This program is free software; you can redistribute it and/or modify it
>> + * under the terms of the GNU General Public License version 2 as published by
>> + * the Free Software Foundation.
>> + *
>> + * Notes
>> + * -----
>> + * The Watchdog consists of a fixed divide-by-4 clock pre-scaler and a 24-bit counter
>> + * which decrements on every clock cycle.
>> + *
>> + */
>> +
>> +#include <linux/clk.h>
>> +#include <linux/io.h>
>> +#include <linux/module.h>
>> +#include <linux/of.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/reboot.h>
>> +#include <linux/watchdog.h>
>> +
>> +/* Registers */
>> +#define LPC_WDT_MOD			0x00
>> +#define LPC_WDT_MOD_WDEN		BIT(0)
>> +#define LPC_WDT_MOD_WDRESET		BIT(1)
>> +#define LPC_WDT_MOD_WDTOF_MASK		0x04
>> +
>> +#define LPC_WDT_TC			0x04
>> +#define LPC_WDT_TC_MIN			0xff
>> +#define LPC_WDT_TC_MAX			0xffffff
>> +
>> +#define LPC_WDT_FEED			0x08
>> +#define LPC_WDT_FEED_MAGIC1		0xaa
>> +#define LPC_WDT_FEED_MAGIC2		0x55
>> +
>> +#define LPC_WDT_TV			0x0c
>> +
>> +/* Clock pre-scaler */
>> +#define LPC_WDT_CLK_DIV			4
>> +
>> +/* Timeout values in seconds */
>> +#define LPC_WDT_DEF_TIMEOUT		1
>> +
> 
> One second ? This is highly unusual. 30 or 60 seconds is more common,
> and one second would be very challenging for user space.
> 
> Any special reason for using such a tight default ?

Considering that LPC18xx Watchdog has a fixed divide-by-4 clock
pre-scaler and a 24-bit counter and that Watchdog clock runs at a fixed
frequency of 12MHz, timeout range goes from 1 to 5 seconds.

I think you're right, 1 sec is very challenging, so it's 5 secs then.

> 
>> +static int heartbeat;
>> +module_param(heartbeat, int, 0);
>> +MODULE_PARM_DESC(heartbeat, "Watchdog heartbeats in seconds "
>> +	"(default=" __MODULE_STRING(LPC_WDT_DEF_TIMEOUT) ")");
>> +
>> +static bool nowayout = WATCHDOG_NOWAYOUT;
>> +module_param(nowayout, bool, 0);
>> +MODULE_PARM_DESC(nowayout, "Watchdog cannot be stopped once started "
>> +	"(default=" __MODULE_STRING(WATCHDOG_NOWAYOUT) ")");
>> +
>> +struct lpc_wdt_dev {
>> +	struct watchdog_device wdt_dev;
>> +	struct clk *wdt_clk;
>> +	struct clk *sys_clk;
>> +	void __iomem *base;
>> +	struct timer_list timer;
>> +	struct notifier_block restart_handler;
>> +};
>> +
>> +static int lpc_wdt_feed(struct watchdog_device *wdt_dev)
>> +{
>> +	unsigned long flags;
>> +	struct lpc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
>> +
> Please order variables length wise where possible.

Okay, I'll do it in next version.

> 
>> +	raw_local_irq_save(flags);
>> +	writel(LPC_WDT_FEED_MAGIC1, wdt->base + LPC_WDT_FEED);
>> +	writel(LPC_WDT_FEED_MAGIC2, wdt->base + LPC_WDT_FEED);
>> +	raw_local_irq_restore(flags);
>> +
>> +	return 0;
>> +}
>> +
>> +static void lpc_wdt_timer_feed(unsigned long data)
>> +{
>> +	unsigned long flags;
>> +	struct watchdog_device *wdt_dev = (struct watchdog_device *) data;
>> +	struct lpc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
>> +
>> +	/*
>> +	 * An abort condition will occur if an interrupt happens during the feed
>> +	 * sequence.
>> +	 */
>> +	raw_local_irq_save(flags);
>> +	writel(LPC_WDT_FEED_MAGIC1, wdt->base + LPC_WDT_FEED);
>> +	writel(LPC_WDT_FEED_MAGIC2, wdt->base + LPC_WDT_FEED);
>> +	raw_local_irq_restore(flags);
>> +
> Unless I am missing something, 
> 	lpc_wdt_feed(wdt_dev);
> should do the same and reduce code duplication.

Totally right. I'll do it in next version.

> 
>> +	/* Use safe value (1/2 of real timeout) */
>> +	mod_timer(&wdt->timer, jiffies + msecs_to_jiffies((wdt_dev->timeout *
>> +							   MSEC_PER_SEC) / 2));
>> +}
>> +
>> +/*
>> + * Since LPC18xx Watchdog cannot be disabled in hardware, we must keep feeding
>> + * it with a timer until userspace watchdog software takes over.
>> + */
>> +static int lpc_wdt_stop(struct watchdog_device *wdt_dev)
>> +{
>> +	lpc_wdt_timer_feed((unsigned long) wdt_dev);
>> +
>> +	return 0;
>> +}
>> +
>> +static void __lpc_wdt_set_timeout(struct lpc_wdt_dev *wdt)
>> +{
>> +	unsigned long clk_rate = clk_get_rate(wdt->wdt_clk);
>> +	unsigned int val;
>> +
>> +	val = DIV_ROUND_UP(wdt->wdt_dev.timeout * clk_rate, LPC_WDT_CLK_DIV);
>> +	writel(val, wdt->base + LPC_WDT_TC);
>> +}
>> +
>> +static int lpc_wdt_set_timeout(struct watchdog_device *wdt_dev,
>> +			       unsigned int new_timeout)
>> +{
>> +	struct lpc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
>> +
>> +	wdt->wdt_dev.timeout = new_timeout;
>> +	__lpc_wdt_set_timeout(wdt);
>> +
>> +	return 0;
>> +}
>> +
>> +
>> +unsigned int lpc_wdt_get_timeleft(struct watchdog_device *wdt_dev)
>> +{
>> +	struct lpc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
>> +	unsigned long clk_rate = clk_get_rate(wdt->wdt_clk);
>> +	unsigned int val;
>> +
>> +	val = readl(wdt->base + LPC_WDT_TV);
>> +	return ((val * LPC_WDT_CLK_DIV) / clk_rate);
>> +}
>> +
>> +static int lpc_wdt_start(struct watchdog_device *wdt_dev)
>> +{
>> +	unsigned int val;
>> +	struct lpc_wdt_dev *wdt = watchdog_get_drvdata(wdt_dev);
>> +
>> +	if (timer_pending(&wdt->timer))
>> +		del_timer(&wdt->timer);
>> +
>> +	val = readl(wdt->base + LPC_WDT_MOD);
>> +	val |= LPC_WDT_MOD_WDEN;
>> +	val |= LPC_WDT_MOD_WDRESET;
>> +	writel(val, wdt->base + LPC_WDT_MOD);
>> +
>> +	/*
>> +	 * Setting the WDEN bit in the WDMOD register is not sufficient to
>> +	 * enable the Watchdog. A valid feed sequence must be completed after
>> +	 * setting WDEN before the Watchdog is capable of generating a reset.
>> +	 */
>> +	lpc_wdt_feed(wdt_dev);
>> +
>> +	return 0;
>> +}
>> +
>> +static struct watchdog_info lpc_wdt_info = {
>> +	.identity	= "LPC 18XX Watchdog",
>> +	.options	= WDIOF_SETTIMEOUT |
>> +			  WDIOF_KEEPALIVEPING |
>> +			  WDIOF_MAGICCLOSE,
>> +};
>> +
>> +static const struct watchdog_ops lpc_wdt_ops = {
>> +	.owner		= THIS_MODULE,
>> +	.start		= lpc_wdt_start,
>> +	.stop		= lpc_wdt_stop,
>> +	.ping		= lpc_wdt_feed,
>> +	.set_timeout	= lpc_wdt_set_timeout,
>> +	.get_timeleft	= lpc_wdt_get_timeleft,
>> +};
>> +
>> +static int lpc_wdt_restart(struct notifier_block *this, unsigned long mode,
>> +			   void *cmd)
>> +{
>> +	int val;
>> +	unsigned long flags;
>> +	struct lpc_wdt_dev *wdt = container_of(this, struct lpc_wdt_dev,
>> +					       restart_handler);
>> +
>> +	/*
>> +	 * Incorrect feed sequence causes immediate watchdog reset if enabled.
>> +	 */
>> +	raw_local_irq_save(flags);
>> +
>> +	val = readl(wdt->base + LPC_WDT_MOD);
>> +	val |= LPC_WDT_MOD_WDEN;
>> +	val |= LPC_WDT_MOD_WDRESET;
>> +	writel(val, wdt->base + LPC_WDT_MOD);
>> +
>> +	writel(LPC_WDT_FEED_MAGIC1, wdt->base + LPC_WDT_FEED);
>> +	writel(LPC_WDT_FEED_MAGIC2, wdt->base + LPC_WDT_FEED);
>> +
>> +	writel(LPC_WDT_FEED_MAGIC1, wdt->base + LPC_WDT_FEED);
>> +	writel(LPC_WDT_FEED_MAGIC1, wdt->base + LPC_WDT_FEED);
>> +
>> +	raw_local_irq_restore(flags);
>> +
>> +	return NOTIFY_OK;
>> +}
>> +
>> +static int lpc_wdt_probe(struct platform_device *pdev)
>> +{
>> +	int ret;
>> +	unsigned long clk_rate;
>> +	struct resource *res;
>> +	struct lpc_wdt_dev *lpc_wdt;
>> +
>> +	lpc_wdt = devm_kzalloc(&pdev->dev, sizeof(*lpc_wdt), GFP_KERNEL);
>> +	if (!lpc_wdt)
>> +		return -ENOMEM;
>> +
>> +	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> +	lpc_wdt->base = devm_ioremap_resource(&pdev->dev, res);
>> +	if (IS_ERR(lpc_wdt->base))
>> +		return PTR_ERR(lpc_wdt->base);
>> +
>> +	lpc_wdt->sys_clk = devm_clk_get(&pdev->dev, "sys");
>> +	if (IS_ERR(lpc_wdt->sys_clk)) {
>> +		dev_err(&pdev->dev, "failed to get the sys clock\n");
>> +		return PTR_ERR(lpc_wdt->sys_clk);
>> +	}
>> +
>> +	lpc_wdt->wdt_clk = devm_clk_get(&pdev->dev, "wdt");
>> +	if (IS_ERR(lpc_wdt->wdt_clk)) {
>> +		dev_err(&pdev->dev, "failed to get the wdt clock\n");
>> +		return PTR_ERR(lpc_wdt->wdt_clk);
>> +	}
>> +
>> +	ret = clk_prepare_enable(lpc_wdt->sys_clk);
>> +	if (ret) {
>> +		dev_err(&pdev->dev, "could not prepare or enable sys clock\n");
>> +		return ret;
>> +	}
>> +
>> +	ret = clk_prepare_enable(lpc_wdt->wdt_clk);
>> +	if (ret) {
>> +		dev_err(&pdev->dev, "could not prepare or enable wdt clock\n");
>> +		goto disable_sys_clk;
>> +	}
>> +
>> +	/* We use the clock rate to calculate timeouts */
>> +	clk_rate = clk_get_rate(lpc_wdt->wdt_clk);
>> +	if (clk_rate == 0) {
>> +		dev_err(&pdev->dev, "failed to get clock rate\n");
>> +		ret = -EINVAL;
>> +		goto disable_wdt_clk;
>> +	}
>> +
>> +	lpc_wdt->wdt_dev.info = &lpc_wdt_info;
>> +	lpc_wdt->wdt_dev.ops = &lpc_wdt_ops;
>> +
>> +	lpc_wdt->wdt_dev.min_timeout =
>> +		DIV_ROUND_UP(LPC_WDT_TC_MIN * LPC_WDT_CLK_DIV, clk_rate);
>> +
>> +	lpc_wdt->wdt_dev.max_timeout =
>> +		(LPC_WDT_TC_MAX * LPC_WDT_CLK_DIV) / clk_rate;
>> +
>> +	lpc_wdt->wdt_dev.parent = &pdev->dev;
>> +	watchdog_set_drvdata(&lpc_wdt->wdt_dev, lpc_wdt);
>> +
>> +	ret = watchdog_init_timeout(&lpc_wdt->wdt_dev, heartbeat, &pdev->dev);
>> +	if (ret < 0)
>> +		lpc_wdt->wdt_dev.timeout = LPC_WDT_DEF_TIMEOUT;
>> +
>> +	__lpc_wdt_set_timeout(lpc_wdt);
>> +
>> +	setup_timer(&lpc_wdt->timer, lpc_wdt_timer_feed,
>> +		    (unsigned long)&lpc_wdt->wdt_dev);
>> +
>> +	watchdog_set_nowayout(&lpc_wdt->wdt_dev, nowayout);
>> +
>> +	platform_set_drvdata(pdev, lpc_wdt);
>> +
>> +	ret = watchdog_register_device(&lpc_wdt->wdt_dev);
>> +	if (ret)
>> +		goto disable_wdt_clk;
>> +
>> +	lpc_wdt->restart_handler.notifier_call = lpc_wdt_restart;
>> +	lpc_wdt->restart_handler.priority = 128;
>> +	ret = register_restart_handler(&lpc_wdt->restart_handler);
>> +	if (ret)
>> +		dev_warn(&pdev->dev, "failed to register restart handler: %d\n",
>> +			 ret);
>> +
>> +	return 0;
>> +
>> +disable_wdt_clk:
>> +	clk_disable_unprepare(lpc_wdt->wdt_clk);
>> +disable_sys_clk:
>> +	clk_disable_unprepare(lpc_wdt->sys_clk);
>> +	return ret;
>> +}
>> +
>> +static void lpc_wdt_shutdown(struct platform_device *pdev)
>> +{
>> +	struct lpc_wdt_dev *lpc_wdt = platform_get_drvdata(pdev);
>> +
>> +	lpc_wdt_stop(&lpc_wdt->wdt_dev);
>> +}
>> +
>> +static int lpc_wdt_remove(struct platform_device *pdev)
>> +{
>> +	struct lpc_wdt_dev *lpc_wdt = platform_get_drvdata(pdev);
>> +
>> +	lpc_wdt_stop(&lpc_wdt->wdt_dev);
> 
> This will keep the timer enabled. It would be interesting to see what
> happens if you build the driver as module and unload it. I think
> it will crash. Can you try ?

Yes, you're right. After module gets unloaded, timer keeps enabled with
a callback that was deallocated from memory.

Since Watchdog cannot be disabled in hardware, it's not going to be
possible to remove the driver without resetting the system. Unless we
could keep a timer set and running outside the module, it won't be
removable.

What do you think?

> 
>> +	watchdog_unregister_device(&lpc_wdt->wdt_dev);
>> +	clk_disable_unprepare(lpc_wdt->wdt_clk);
>> +	clk_disable_unprepare(lpc_wdt->sys_clk);
>> +
>> +	return 0;
>> +}
>> +
>> +static const struct of_device_id lpc_wdt_match[] = {
>> +	{ .compatible = "nxp,lpc18xx-wdt" },
>> +	{}
>> +};
>> +MODULE_DEVICE_TABLE(of, lpc_wdt_match);
>> +
>> +static struct platform_driver lpc_wdt_driver = {
>> +	.driver = {
>> +		.name = "lpc18xx-wdt",
>> +		.of_match_table	= lpc_wdt_match,
>> +	},
>> +	.probe = lpc_wdt_probe,
>> +	.remove = lpc_wdt_remove,
>> +	.shutdown = lpc_wdt_shutdown,
>> +};
>> +module_platform_driver(lpc_wdt_driver);
>> +
>> +MODULE_AUTHOR("Ariel D'Alessandro <ariel@vanguardiasur.com.ar>");
>> +MODULE_DESCRIPTION("NXP LPC18xx Windowed Watchdog Timer Driver");
>> +MODULE_LICENSE("GPL v2");
>> -- 
>> 1.9.1
>>

  reply	other threads:[~2015-06-28 18:13 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-06-26 18:24 [PATCH 0/2] Add support for NXP LPC18xx Watchdog timer Ariel D'Alessandro
2015-06-26 18:24 ` [PATCH 1/2] watchdog: NXP LPC18XX Windowed Watchdog Timer Driver Ariel D'Alessandro
2015-06-26 19:07   ` Guenter Roeck
2015-06-28 18:13     ` Ariel D'Alessandro [this message]
2015-06-29  4:47       ` Guenter Roeck
2015-07-01 11:30         ` adalessandro
2015-07-01 12:02           ` Ariel D'Alessandro
2015-07-01 13:54             ` Guenter Roeck
2015-07-01 14:38               ` Wim Van Sebroeck
2015-07-01 15:22               ` Ezequiel Garcia
2015-07-01 15:22                 ` Ezequiel Garcia
2015-07-01 17:03                 ` Guenter Roeck
2015-07-01 17:03                   ` Guenter Roeck
2015-06-26 18:24 ` [PATCH 2/2] DT: watchdog: Add NXP LPC18XX Watchdog Timer binding documentation Ariel D'Alessandro

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=55903943.10309@vanguardiasur.com.ar \
    --to=ariel@vanguardiasur.com.ar \
    --cc=ezequiel@vanguardiasur.com.ar \
    --cc=linux-watchdog@vger.kernel.org \
    --cc=linux@roeck-us.net \
    --cc=wim@iguana.be \
    /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.