Re: [PATCH v3 1/2] watchdog: add Intel MID watchdog driver support

From: David Cohen
Date: Tue Apr 15 2014 - 15:24:30 EST

Next message: Vince Weaver: "Re: [bisected] broken make-kpkg kernel build in 3.15-rc1"
Previous message: Pascal COMBES: "[PATCH v3 3/3] Staging: crystalhd: Improve kzalloc calls in crystalhd_lnx.c"
In reply to: Guenter Roeck: "Re: [PATCH v3 1/2] watchdog: add Intel MID watchdog driver support"
Next in thread: David Cohen: "[PATCH v4 0/2] Initial implementation of Intel MID watchdog driver"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Tue, Apr 15, 2014 at 12:01:49PM -0700, Guenter Roeck wrote:
> On Tue, Apr 15, 2014 at 11:41:11AM -0700, David Cohen wrote:
> > Add initial Intel MID watchdog driver support.
> >
> > This driver is an initial implementation of generic Intel MID watchdog
> > driver. Currently it supports Intel Merrifield platform.
> >
> > Signed-off-by: Eric Ernst <eric.ernst@xxxxxxxxx>
> > Signed-off-by: David Cohen <david.a.cohen@xxxxxxxxxxxxxxx>
> > ---
> > drivers/watchdog/Kconfig | 13 +++
> > drivers/watchdog/Makefile | 1 +
> > drivers/watchdog/intel-mid_wdt.c | 187 +++++++++++++++++++++++++++++++++++++++
> > include/linux/intel-mid_wdt.h | 20 +++++
>
> include/linux/platform_data/intel-mid_wdt.h ?

Ack

>
> > 4 files changed, 221 insertions(+)
> > create mode 100644 drivers/watchdog/intel-mid_wdt.c
> > create mode 100644 include/linux/intel-mid_wdt.h
> >
> > diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
> > index 79d25894343a..0e110c37aa5c 100644
> > --- a/drivers/watchdog/Kconfig
> > +++ b/drivers/watchdog/Kconfig
> > @@ -643,6 +643,19 @@ config INTEL_SCU_WATCHDOG
> >
> > To compile this driver as a module, choose M here.
> >
> > +config INTEL_MID_WATCHDOG
> > + tristate "Intel MID Watchdog Timer"
> > + depends on X86_INTEL_MID
> > + select WATCHDOG_CORE
> > + ---help---
> > + Watchdog timer driver built into the Intel SCU for Intel MID
> > + Platforms.
> > +
> > + This driver currently supports only the watchdog evolution
> > + implementation in SCU, available for Merrifield generation.
> > +
> > + To compile this driver as a module, choose M here.
> > +
> > config ITCO_WDT
> > tristate "Intel TCO Timer/Watchdog"
> > depends on (X86 || IA64) && PCI
> > diff --git a/drivers/watchdog/Makefile b/drivers/watchdog/Makefile
> > index 985a66cda76f..b927e7496eb4 100644
> > --- a/drivers/watchdog/Makefile
> > +++ b/drivers/watchdog/Makefile
> > @@ -113,6 +113,7 @@ obj-$(CONFIG_W83977F_WDT) += w83977f_wdt.o
> > obj-$(CONFIG_MACHZ_WDT) += machzwd.o
> > obj-$(CONFIG_SBC_EPX_C3_WATCHDOG) += sbc_epx_c3.o
> > obj-$(CONFIG_INTEL_SCU_WATCHDOG) += intel_scu_watchdog.o
> > +obj-$(CONFIG_INTEL_MID_WATCHDOG) += intel-mid_wdt.o
> >
> > # M32R Architecture
> >
> > diff --git a/drivers/watchdog/intel-mid_wdt.c b/drivers/watchdog/intel-mid_wdt.c
> > new file mode 100644
> > index 000000000000..d8bf3c3cd2b7
> > --- /dev/null
> > +++ b/drivers/watchdog/intel-mid_wdt.c
> > @@ -0,0 +1,187 @@
> > +/*
> > + * intel-mid_wdt: generic Intel MID SCU watchdog driver
> > + *
> > + * Platforms supported so far:
> > + * - Merrifield only
> > + *
> > + * Copyright (C) 2014 Intel Corporation. All rights reserved.
> > + * Contact: David Cohen <david.a.cohen@xxxxxxxxxxxxxxx>
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of version 2 of the GNU General
> > + * Public License as published by the Free Software Foundation.
> > + */
> > +
> > +#include <linux/intel-mid_wdt.h>
> > +#include <linux/interrupt.h>
> > +#include <linux/module.h>
> > +#include <linux/nmi.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/watchdog.h>
> > +
> > +#include <asm/intel_scu_ipc.h>
> > +#include <asm/intel-mid.h>
> > +
> > +#define IPC_WATCHDOG 0xf8
> > +
> > +#define MID_WDT_PRETIMEOUT 15
> > +#define MID_WDT_TIMEOUT_MIN (1 + MID_WDT_PRETIMEOUT)
> > +#define MID_WDT_TIMEOUT_MAX 170
> > +#define MID_WDT_DEFAULT_TIMEOUT 90
> > +
> > +/* SCU watchdog messages */
> > +enum {
> > + SCU_WATCHDOG_START = 0,
> > + SCU_WATCHDOG_STOP,
> > + SCU_WATCHDOG_KEEPALIVE,
> > +};
> > +
> > +static inline int wdt_command(int sub, u32 *in, int inlen)
> > +{
> > + return intel_scu_ipc_command(IPC_WATCHDOG, sub, in, inlen, NULL, 0);
> > +}
> > +
> > +static int wdt_start(struct watchdog_device *wd)
> > +{
> > + int ret, in_size;
> > + int timeout = wd->timeout;
> > + struct ipc_wd_start {
> > + u32 pretimeout;
> > + u32 timeout;
> > + } ipc_wd_start = { timeout - MID_WDT_PRETIMEOUT, timeout };
> > +
> > + /*
> > + * SCU expects the input size for watchdog IPC to
> > + * be based on double-word
> > + */
> > + in_size = (sizeof(ipc_wd_start) + 3) / 4;
> > +
> DIV_ROUND_UP ?

Ack

>
> Not sure how "double-word" matches the calculation. What is a "double-word" ?
> 4 bytes ?
>
> > + ret = wdt_command(SCU_WATCHDOG_START, (u32 *)&ipc_wd_start, in_size);
> > + if (ret) {
> > + struct device *dev = watchdog_get_drvdata(wd);
> > + dev_crit(dev, "error starting watchdog: %d\n", ret);
> > + return -EIO;
>
> I am quite sure sparse is going to complain about this ...
> Why not just return ret ? This would result in a more detailed
> error code/reason.

wdt_command() may return errors != -EIO. But IMHO -EIO makes more sense
whatever is the issue regarding to IPC in this scenario.
But I can return ret as per request. Same to cases below.

>
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static int wdt_ping(struct watchdog_device *wd)
> > +{
> > + int ret;
> > +
> > + ret = wdt_command(SCU_WATCHDOG_KEEPALIVE, NULL, 0);
> > + if (ret) {
> > + struct device *dev = watchdog_get_drvdata(wd);
> > + dev_crit(dev, "Error executing keepalive: 0x%x\n", ret);
> > + return -EIO;
>
> Same here.
>
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static int wdt_stop(struct watchdog_device *wd)
> > +{
> > + int ret;
> > +
> > + ret = wdt_command(SCU_WATCHDOG_STOP, NULL, 0);
> > + if (ret) {
> > + struct device *dev = watchdog_get_drvdata(wd);
> > + dev_crit(dev, "Error stopping watchdog: 0x%x\n", ret);
> > + return -EIO;
>
> Same here.
>
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static irqreturn_t mid_wdt_warning_irq(int irq, void *dev_id)
>
> This isn't really a warning, so the function name is a bit misleading.

Ack

>
> > +{
> > + panic("Kernel Watchdog");
> > +
> > + /* This code should not be reached */
> > + return IRQ_HANDLED;
> > +}
> > +
> > +static const struct watchdog_info mid_wdt_info = {
> > + .identity = "Intel MID SCU watchdog",
> > + .options = WDIOF_KEEPALIVEPING | WDIOF_SETTIMEOUT,
> > +};
> > +
> > +static const struct watchdog_ops mid_wdt_ops = {
> > + .owner = THIS_MODULE,
> > + .start = wdt_start,
> > + .stop = wdt_stop,
> > + .ping = wdt_ping,
> > +};
> > +
> > +static int mid_wdt_probe(struct platform_device *pdev)
> > +{
> > + struct watchdog_device *wdt_dev;
> > + struct intel_mid_wdt_pdata *pdata = pdev->dev.platform_data;
> > + int ret;
> > +
> > + if (!pdata) {
> > + dev_err(&pdev->dev, "missing platform data\n");
> > + return -EINVAL;
> > + }
> > +
> > + if (pdata->probe) {
> > + ret = pdata->probe();
>
> Sure you don't want to pass the platform device as parameter ?
> May not be needed today, but it might be better than to rely on the assumption
> that the called code gets it from a static variable or structure (if needed).
>
> This would enable you, for example, to use dev_warn() instead of pr_warn()
> in the platform code.

Ack.

>
> > + if (ret)
> > + return ret;
> > + }
> > +
> > + wdt_dev = devm_kzalloc(&pdev->dev, sizeof(*wdt_dev), GFP_KERNEL);
> > + if (!wdt_dev)
> > + return -ENOMEM;
> > +
> > + wdt_dev->info = &mid_wdt_info;
> > + wdt_dev->ops = &mid_wdt_ops;
> > + wdt_dev->min_timeout = MID_WDT_TIMEOUT_MIN;
> > + wdt_dev->max_timeout = MID_WDT_TIMEOUT_MAX;
> > + wdt_dev->timeout = MID_WDT_DEFAULT_TIMEOUT;
> > +
> > + watchdog_set_drvdata(wdt_dev, &pdev->dev);
> > + platform_set_drvdata(pdev, wdt_dev);
> > +
> > + ret = devm_request_irq(&pdev->dev, pdata->irq, mid_wdt_warning_irq,
> > + IRQF_SHARED | IRQF_NO_SUSPEND, "watchdog",
> > + wdt_dev);
>
> If irq is mandatory, it might make sense to check if it is valid.

I can add this extra sanity check.

Thanks for your review.

BR, David

>
> > + if (ret) {
> > + dev_err(&pdev->dev, "error requesting warning irq %d\n",
> > + pdata->irq);
> > + return ret;
> > + }
> > +
> > + ret = watchdog_register_device(wdt_dev);
> > + if (ret) {
> > + dev_err(&pdev->dev, "error registering watchdog device\n");
> > + return ret;
> > + }
> > +
> > + dev_info(&pdev->dev, "Intel MID watchdog device probed\n");
> > +
> > + return 0;
> > +}
> > +
> > +static int mid_wdt_remove(struct platform_device *pdev)
> > +{
> > + struct watchdog_device *wd = platform_get_drvdata(pdev);
> > + watchdog_unregister_device(wd);
> > + return 0;
> > +}
> > +
> > +static struct platform_driver mid_wdt_driver = {
> > + .probe = mid_wdt_probe,
> > + .remove = mid_wdt_remove,
> > + .driver = {
> > + .owner = THIS_MODULE,
> > + .name = "intel_mid_wdt",
> > + },
> > +};
> > +
> > +module_platform_driver(mid_wdt_driver);
> > +
> > +MODULE_AUTHOR("David Cohen <david.a.cohen@xxxxxxxxxxxxxxx>");
> > +MODULE_DESCRIPTION("Watchdog Driver for Intel MID platform");
> > +MODULE_LICENSE("GPL");
> > diff --git a/include/linux/intel-mid_wdt.h b/include/linux/intel-mid_wdt.h
> > new file mode 100644
> > index 000000000000..705dff1c62f8
> > --- /dev/null
> > +++ b/include/linux/intel-mid_wdt.h
> > @@ -0,0 +1,20 @@
> > +/*
> > + * intel-mid_wdt: generic Intel MID SCU watchdog driver
> > + *
> > + * Copyright (C) 2014 Intel Corporation. All rights reserved.
> > + * Contact: David Cohen <david.a.cohen@xxxxxxxxxxxxxxx>
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of version 2 of the GNU General
> > + * Public License as published by the Free Software Foundation.
> > + */
> > +
> > +#ifndef __INTEL_MID_WDT_H__
> > +#define __INTEL_MID_WDT_H__
> > +
> > +struct intel_mid_wdt_pdata {
> > + int irq;
> > + int (*probe)(void);
> > +};
> > +
> > +#endif /*__INTEL_MID_WDT_H__*/
> > --
> > 1.9.1
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-watchdog" in
> > the body of a message to majordomo@xxxxxxxxxxxxxxx
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> >
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Next message: Vince Weaver: "Re: [bisected] broken make-kpkg kernel build in 3.15-rc1"
Previous message: Pascal COMBES: "[PATCH v3 3/3] Staging: crystalhd: Improve kzalloc calls in crystalhd_lnx.c"
In reply to: Guenter Roeck: "Re: [PATCH v3 1/2] watchdog: add Intel MID watchdog driver support"
Next in thread: David Cohen: "[PATCH v4 0/2] Initial implementation of Intel MID watchdog driver"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]