| Message ID | 20220801092150.4449-1-pboos@baylibre.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] watchdog: meson: keep running if already active | expand |
On 8/1/22 02:21, Philippe Boos wrote: > If the watchdog is already running (e.g.: started by bootloader) then > the kernel driver should keep the watchdog active but the amlogic driver > turns it off. > > Let the driver fix the clock rate if already active because we do not > know the previous timebase value. To avoid unintentional resetting we > temporarily set it to its maximum value. > > Then keep the enable bit if is was previously active. > > Signed-off-by: Philippe Boos <pboos@baylibre.com> What changed since v1 ? Guenter > --- > drivers/watchdog/meson_gxbb_wdt.c | 24 +++++++++++++++++++----- > 1 file changed, 19 insertions(+), 5 deletions(-) > > diff --git a/drivers/watchdog/meson_gxbb_wdt.c b/drivers/watchdog/meson_gxbb_wdt.c > index 5a9ca10fbcfa..8be8fd9e5637 100644 > --- a/drivers/watchdog/meson_gxbb_wdt.c > +++ b/drivers/watchdog/meson_gxbb_wdt.c > @@ -146,6 +146,7 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) > struct device *dev = &pdev->dev; > struct meson_gxbb_wdt *data; > int ret; > + u32 ctrl_reg; > > data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); > if (!data) > @@ -177,13 +178,26 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) > data->wdt_dev.timeout = DEFAULT_TIMEOUT; > watchdog_set_drvdata(&data->wdt_dev, data); > > + ctrl_reg = readl(data->reg_base + GXBB_WDT_CTRL_REG) & > + GXBB_WDT_CTRL_EN; > + > + if (ctrl_reg) { > + /* Watchdog is running - keep it running but extend timeout > + * to the maximum while setting the timebase > + */ > + set_bit(WDOG_HW_RUNNING, &data->wdt_dev.status); > + meson_gxbb_wdt_set_timeout(&data->wdt_dev, > + GXBB_WDT_TCNT_SETUP_MASK / 1000); > + } > + > /* Setup with 1ms timebase */ > - writel(((clk_get_rate(data->clk) / 1000) & GXBB_WDT_CTRL_DIV_MASK) | > - GXBB_WDT_CTRL_EE_RESET | > - GXBB_WDT_CTRL_CLK_EN | > - GXBB_WDT_CTRL_CLKDIV_EN, > - data->reg_base + GXBB_WDT_CTRL_REG); > + ctrl_reg |= ((clk_get_rate(data->clk) / 1000) & > + GXBB_WDT_CTRL_DIV_MASK) | > + GXBB_WDT_CTRL_EE_RESET | > + GXBB_WDT_CTRL_CLK_EN | > + GXBB_WDT_CTRL_CLKDIV_EN; > > + writel(ctrl_reg, data->reg_base + GXBB_WDT_CTRL_REG); > meson_gxbb_wdt_set_timeout(&data->wdt_dev, data->wdt_dev.timeout); > > watchdog_stop_on_reboot(&data->wdt_dev);
On 8/8/22 13:58, Guenter Roeck wrote: > On 8/1/22 02:21, Philippe Boos wrote: >> If the watchdog is already running (e.g.: started by bootloader) then >> the kernel driver should keep the watchdog active but the amlogic driver >> turns it off. >> >> Let the driver fix the clock rate if already active because we do not >> know the previous timebase value. To avoid unintentional resetting we >> temporarily set it to its maximum value. >> >> Then keep the enable bit if is was previously active. >> >> Signed-off-by: Philippe Boos <pboos@baylibre.com> > > What changed since v1 ? In v1, we read watchdog's enable bit before writing its config register, then we write the register, this will always stop the watchdog. If it was previously active we restart it. So, in v1, if the kernel crashes just before the watchdog restarts it will be stuck forever. In v2, we read watchdog's enable bit before writing its config register, if it is already active we keep the enable bit when writing the register. Regards, Philippe Boos > > Guenter > >> --- >> drivers/watchdog/meson_gxbb_wdt.c | 24 +++++++++++++++++++----- >> 1 file changed, 19 insertions(+), 5 deletions(-) >> >> diff --git a/drivers/watchdog/meson_gxbb_wdt.c b/drivers/watchdog/meson_gxbb_wdt.c >> index 5a9ca10fbcfa..8be8fd9e5637 100644 >> --- a/drivers/watchdog/meson_gxbb_wdt.c >> +++ b/drivers/watchdog/meson_gxbb_wdt.c >> @@ -146,6 +146,7 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) >> struct device *dev = &pdev->dev; >> struct meson_gxbb_wdt *data; >> int ret; >> + u32 ctrl_reg; >> data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); >> if (!data) >> @@ -177,13 +178,26 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) >> data->wdt_dev.timeout = DEFAULT_TIMEOUT; >> watchdog_set_drvdata(&data->wdt_dev, data); >> + ctrl_reg = readl(data->reg_base + GXBB_WDT_CTRL_REG) & >> + GXBB_WDT_CTRL_EN; >> + >> + if (ctrl_reg) { >> + /* Watchdog is running - keep it running but extend timeout >> + * to the maximum while setting the timebase >> + */ >> + set_bit(WDOG_HW_RUNNING, &data->wdt_dev.status); >> + meson_gxbb_wdt_set_timeout(&data->wdt_dev, >> + GXBB_WDT_TCNT_SETUP_MASK / 1000); >> + } >> + >> /* Setup with 1ms timebase */ >> - writel(((clk_get_rate(data->clk) / 1000) & GXBB_WDT_CTRL_DIV_MASK) | >> - GXBB_WDT_CTRL_EE_RESET | >> - GXBB_WDT_CTRL_CLK_EN | >> - GXBB_WDT_CTRL_CLKDIV_EN, >> - data->reg_base + GXBB_WDT_CTRL_REG); >> + ctrl_reg |= ((clk_get_rate(data->clk) / 1000) & >> + GXBB_WDT_CTRL_DIV_MASK) | >> + GXBB_WDT_CTRL_EE_RESET | >> + GXBB_WDT_CTRL_CLK_EN | >> + GXBB_WDT_CTRL_CLKDIV_EN; >> + writel(ctrl_reg, data->reg_base + GXBB_WDT_CTRL_REG); >> meson_gxbb_wdt_set_timeout(&data->wdt_dev, data->wdt_dev.timeout); >> watchdog_stop_on_reboot(&data->wdt_dev); >
On Mon 01 Aug 2022 at 11:21, Philippe Boos <pboos@baylibre.com> wrote: > If the watchdog is already running (e.g.: started by bootloader) then > the kernel driver should keep the watchdog active but the amlogic driver > turns it off. > > Let the driver fix the clock rate if already active because we do not > know the previous timebase value. To avoid unintentional resetting we > temporarily set it to its maximum value. > > Then keep the enable bit if is was previously active. > > Signed-off-by: Philippe Boos <pboos@baylibre.com> Reviewed-by: Jerome Brunet <jbrunet@baylibre.com>
On Mon, Aug 01, 2022 at 11:21:50AM +0200, Philippe Boos wrote: > If the watchdog is already running (e.g.: started by bootloader) then > the kernel driver should keep the watchdog active but the amlogic driver > turns it off. > > Let the driver fix the clock rate if already active because we do not > know the previous timebase value. To avoid unintentional resetting we > temporarily set it to its maximum value. > > Then keep the enable bit if is was previously active. > > Signed-off-by: Philippe Boos <pboos@baylibre.com> > Reviewed-by: Jerome Brunet <jbrunet@baylibre.com> Reviewed-by: Guenter Roeck <linux@roeck-us.net> > --- > drivers/watchdog/meson_gxbb_wdt.c | 24 +++++++++++++++++++----- > 1 file changed, 19 insertions(+), 5 deletions(-) > > diff --git a/drivers/watchdog/meson_gxbb_wdt.c b/drivers/watchdog/meson_gxbb_wdt.c > index 5a9ca10fbcfa..8be8fd9e5637 100644 > --- a/drivers/watchdog/meson_gxbb_wdt.c > +++ b/drivers/watchdog/meson_gxbb_wdt.c > @@ -146,6 +146,7 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) > struct device *dev = &pdev->dev; > struct meson_gxbb_wdt *data; > int ret; > + u32 ctrl_reg; > > data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); > if (!data) > @@ -177,13 +178,26 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) > data->wdt_dev.timeout = DEFAULT_TIMEOUT; > watchdog_set_drvdata(&data->wdt_dev, data); > > + ctrl_reg = readl(data->reg_base + GXBB_WDT_CTRL_REG) & > + GXBB_WDT_CTRL_EN; > + > + if (ctrl_reg) { > + /* Watchdog is running - keep it running but extend timeout > + * to the maximum while setting the timebase > + */ > + set_bit(WDOG_HW_RUNNING, &data->wdt_dev.status); > + meson_gxbb_wdt_set_timeout(&data->wdt_dev, > + GXBB_WDT_TCNT_SETUP_MASK / 1000); > + } > + > /* Setup with 1ms timebase */ > - writel(((clk_get_rate(data->clk) / 1000) & GXBB_WDT_CTRL_DIV_MASK) | > - GXBB_WDT_CTRL_EE_RESET | > - GXBB_WDT_CTRL_CLK_EN | > - GXBB_WDT_CTRL_CLKDIV_EN, > - data->reg_base + GXBB_WDT_CTRL_REG); > + ctrl_reg |= ((clk_get_rate(data->clk) / 1000) & > + GXBB_WDT_CTRL_DIV_MASK) | > + GXBB_WDT_CTRL_EE_RESET | > + GXBB_WDT_CTRL_CLK_EN | > + GXBB_WDT_CTRL_CLKDIV_EN; > > + writel(ctrl_reg, data->reg_base + GXBB_WDT_CTRL_REG); > meson_gxbb_wdt_set_timeout(&data->wdt_dev, data->wdt_dev.timeout); > > watchdog_stop_on_reboot(&data->wdt_dev);
diff --git a/drivers/watchdog/meson_gxbb_wdt.c b/drivers/watchdog/meson_gxbb_wdt.c index 5a9ca10fbcfa..8be8fd9e5637 100644 --- a/drivers/watchdog/meson_gxbb_wdt.c +++ b/drivers/watchdog/meson_gxbb_wdt.c @@ -146,6 +146,7 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) struct device *dev = &pdev->dev; struct meson_gxbb_wdt *data; int ret; + u32 ctrl_reg; data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL); if (!data) @@ -177,13 +178,26 @@ static int meson_gxbb_wdt_probe(struct platform_device *pdev) data->wdt_dev.timeout = DEFAULT_TIMEOUT; watchdog_set_drvdata(&data->wdt_dev, data); + ctrl_reg = readl(data->reg_base + GXBB_WDT_CTRL_REG) & + GXBB_WDT_CTRL_EN; + + if (ctrl_reg) { + /* Watchdog is running - keep it running but extend timeout + * to the maximum while setting the timebase + */ + set_bit(WDOG_HW_RUNNING, &data->wdt_dev.status); + meson_gxbb_wdt_set_timeout(&data->wdt_dev, + GXBB_WDT_TCNT_SETUP_MASK / 1000); + } + /* Setup with 1ms timebase */ - writel(((clk_get_rate(data->clk) / 1000) & GXBB_WDT_CTRL_DIV_MASK) | - GXBB_WDT_CTRL_EE_RESET | - GXBB_WDT_CTRL_CLK_EN | - GXBB_WDT_CTRL_CLKDIV_EN, - data->reg_base + GXBB_WDT_CTRL_REG); + ctrl_reg |= ((clk_get_rate(data->clk) / 1000) & + GXBB_WDT_CTRL_DIV_MASK) | + GXBB_WDT_CTRL_EE_RESET | + GXBB_WDT_CTRL_CLK_EN | + GXBB_WDT_CTRL_CLKDIV_EN; + writel(ctrl_reg, data->reg_base + GXBB_WDT_CTRL_REG); meson_gxbb_wdt_set_timeout(&data->wdt_dev, data->wdt_dev.timeout); watchdog_stop_on_reboot(&data->wdt_dev);
If the watchdog is already running (e.g.: started by bootloader) then the kernel driver should keep the watchdog active but the amlogic driver turns it off. Let the driver fix the clock rate if already active because we do not know the previous timebase value. To avoid unintentional resetting we temporarily set it to its maximum value. Then keep the enable bit if is was previously active. Signed-off-by: Philippe Boos <pboos@baylibre.com> --- drivers/watchdog/meson_gxbb_wdt.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-)