diff mbox series

watchdog: rza_wdt: Extend clock sources for RZ/A2

Message ID 20180905173743.47938-1-chris.brandt@renesas.com (mailing list archive)
State Superseded
Delegated to: Geert Uytterhoeven
Headers show
Series watchdog: rza_wdt: Extend clock sources for RZ/A2 | expand

Commit Message

Chris Brandt Sept. 5, 2018, 5:37 p.m. UTC
The RZ/A2 watchdog timer extends the clock source options in order to give
you longer timeout options.

Signed-off-by: Chris Brandt <chris.brandt@renesas.com>
---
 drivers/watchdog/rza_wdt.c | 75 ++++++++++++++++++++++++++++++++++++----------
 1 file changed, 59 insertions(+), 16 deletions(-)

Comments

Geert Uytterhoeven Sept. 6, 2018, 10:07 a.m. UTC | #1
Hi Chris,

On Wed, Sep 5, 2018 at 7:40 PM Chris Brandt <chris.brandt@renesas.com> wrote:
> The RZ/A2 watchdog timer extends the clock source options in order to give
> you longer timeout options.
>
> Signed-off-by: Chris Brandt <chris.brandt@renesas.com>

Thanks for your patch!

> --- a/drivers/watchdog/rza_wdt.c
> +++ b/drivers/watchdog/rza_wdt.c
> @@ -38,8 +38,37 @@ struct rza_wdt {
>         struct watchdog_device wdev;
>         void __iomem *base;
>         struct clk *clk;
> +       bool ext_cks;
> +       u8 count;
> +       u8 cks;
> +       u8 timeout;
>  };
>
> +static void rza_wdt_calc_timeout(struct rza_wdt *priv, int timeout)
> +{
> +       int rate = clk_get_rate(priv->clk);
> +       u16 counter;
> +
> +       if (priv->ext_cks) {
> +               /* Add the 1 to deal with the inevitable fraction */
> +               counter = ((timeout * rate) /  4194304) + 1;

"timeout * rate" may overflow
DIV_ROUND_UP()?

> +               if (counter > 255)
> +                       counter = 0;
> +
> +               priv->cks = 0xF;
> +               priv->count = 256 - counter;
> +       } else {
> +               /* Start timer with longest timeout */
> +               priv->cks = 7;
> +               priv->count = 0;
> +       }
> +
> +       priv->timeout = timeout;
> +
> +       pr_debug("%s: timeout set to %d (WTCNT=%d)\n", __func__,

%u for unsigned.

> +                timeout, priv->count);
> +}

> @@ -150,20 +184,29 @@ static int rza_wdt_probe(struct platform_device *pdev)
>                 return -ENOENT;
>         }
>
> -       /* Assume slowest clock rate possible (CKS=7) */
> -       rate /= 16384;
> +       if (of_device_is_compatible(np, "renesas,r7s9210-wdt"))
> +               priv->ext_cks = true;

That's not the proper way to support multiple devices.
Please add an entry for "renesas,r7s9210-wdt" to rza_wdt_of_match[].
"renesas,r7s9210-wdt" must be documented in the DT bindings.

I suggest storing cks in rza_wdt_of_match[].data, and retrieving it with
of_device_get_match_data() in your probe function, so you can use that to
differentiate, and get rid of the ext_cks flag.

BTW, is the RZ/A2 WDT compatible with the RZ/A1 WDT, i.e. does it work
with the unmodified driver? If not, "renesas,rza-wdt" must not be used as a
fallback.

Gr{oetje,eeting}s,

                        Geert
Chris Brandt Sept. 6, 2018, 12:44 p.m. UTC | #2
Hi Geert,

On Thursday, September 06, 2018, Geert Uytterhoeven wrote:
> Thanks for your patch!

Thank you for your review.

> "timeout * rate" may overflow
> DIV_ROUND_UP()?
OK.

> > +       pr_debug("%s: timeout set to %d (WTCNT=%d)\n", __func__,
> 
> %u for unsigned.
OK.


> > +       if (of_device_is_compatible(np, "renesas,r7s9210-wdt"))
> > +               priv->ext_cks = true;
> 
> That's not the proper way to support multiple devices.
> Please add an entry for "renesas,r7s9210-wdt" to rza_wdt_of_match[].
> "renesas,r7s9210-wdt" must be documented in the DT bindings.
> 
> I suggest storing cks in rza_wdt_of_match[].data, and retrieving it with
> of_device_get_match_data() in your probe function, so you can use that to
> differentiate, 

OK, I will change to that.

> and get rid of the ext_cks flag.

I still need to keep track if it's a RZ/A1 or RZ/A2 (ext_cks) for 
functions outside of probe (rza_wdt_calc_timeout).

> BTW, is the RZ/A2 WDT compatible with the RZ/A1 WDT, i.e. does it work
> with the unmodified driver? If not, "renesas,rza-wdt" must not be used as
> a
> fallback.

I would say it is not, because the dividers are different (meaning the 
calculated timeouts would not be correct).

Thanks,
Chris
Geert Uytterhoeven Sept. 6, 2018, 12:48 p.m. UTC | #3
Hi Chris,

On Thu, Sep 6, 2018 at 2:44 PM Chris Brandt <Chris.Brandt@renesas.com> wrote:
> On Thursday, September 06, 2018, Geert Uytterhoeven wrote:
> > > +       if (of_device_is_compatible(np, "renesas,r7s9210-wdt"))
> > > +               priv->ext_cks = true;
> >
> > That's not the proper way to support multiple devices.
> > Please add an entry for "renesas,r7s9210-wdt" to rza_wdt_of_match[].
> > "renesas,r7s9210-wdt" must be documented in the DT bindings.
> >
> > I suggest storing cks in rza_wdt_of_match[].data, and retrieving it with
> > of_device_get_match_data() in your probe function, so you can use that to
> > differentiate,
>
> OK, I will change to that.
>
> > and get rid of the ext_cks flag.
>
> I still need to keep track if it's a RZ/A1 or RZ/A2 (ext_cks) for
> functions outside of probe (rza_wdt_calc_timeout).

You can check if priv->cks > 7?

Gr{oetje,eeting}s,

                        Geert
Chris Brandt Sept. 6, 2018, 12:51 p.m. UTC | #4
Hi Geert,

On Thursday, September 06, 2018, Geert Uytterhoeven wrote:
> > > and get rid of the ext_cks flag.
> >
> > I still need to keep track if it's a RZ/A1 or RZ/A2 (ext_cks) for
> > functions outside of probe (rza_wdt_calc_timeout).
> 
> You can check if priv->cks > 7?

That's a good point! I'll do that.

Thanks,
Chris
Chris Brandt Sept. 6, 2018, 4:11 p.m. UTC | #5
Hi Geert,

Correction:

On Thursday, September 06, 2018, Chris Brandt wrote:
> > BTW, is the RZ/A2 WDT compatible with the RZ/A1 WDT, i.e. does it work
> > with the unmodified driver? If not, "renesas,rza-wdt" must not be used
> as
> > a
> > fallback.
> 
> I would say it is not, because the dividers are different (meaning the
> calculated timeouts would not be correct).

Actually, the clock dividers are the same for the first three bits 
between RZ/A1 and RZ/A2. So you could use "renesas,rza-wdt" with an RZ/A2 and
it would work exactly the same as RZ/A1.


Chris
diff mbox series

Patch

diff --git a/drivers/watchdog/rza_wdt.c b/drivers/watchdog/rza_wdt.c
index e618218d2374..78e0d61519ec 100644
--- a/drivers/watchdog/rza_wdt.c
+++ b/drivers/watchdog/rza_wdt.c
@@ -38,8 +38,37 @@  struct rza_wdt {
 	struct watchdog_device wdev;
 	void __iomem *base;
 	struct clk *clk;
+	bool ext_cks;
+	u8 count;
+	u8 cks;
+	u8 timeout;
 };
 
+static void rza_wdt_calc_timeout(struct rza_wdt *priv, int timeout)
+{
+	int rate = clk_get_rate(priv->clk);
+	u16 counter;
+
+	if (priv->ext_cks) {
+		/* Add the 1 to deal with the inevitable fraction */
+		counter = ((timeout * rate) /  4194304) + 1;
+		if (counter > 255)
+			counter = 0;
+
+		priv->cks = 0xF;
+		priv->count = 256 - counter;
+	} else {
+		/* Start timer with longest timeout */
+		priv->cks = 7;
+		priv->count = 0;
+	}
+
+	priv->timeout = timeout;
+
+	pr_debug("%s: timeout set to %d (WTCNT=%d)\n", __func__,
+		 timeout, priv->count);
+}
+
 static int rza_wdt_start(struct watchdog_device *wdev)
 {
 	struct rza_wdt *priv = watchdog_get_drvdata(wdev);
@@ -51,13 +80,12 @@  static int rza_wdt_start(struct watchdog_device *wdev)
 	readb(priv->base + WRCSR);
 	writew(WRCSR_CLEAR_WOVF, priv->base + WRCSR);
 
-	/*
-	 * Start timer with slowest clock source and reset option enabled.
-	 */
+	rza_wdt_calc_timeout(priv, wdev->timeout);
+
 	writew(WRCSR_MAGIC | WRCSR_RSTE, priv->base + WRCSR);
-	writew(WTCNT_MAGIC | 0, priv->base + WTCNT);
-	writew(WTCSR_MAGIC | WTSCR_WT | WTSCR_TME | WTSCR_CKS(7),
-	       priv->base + WTCSR);
+	writew(WTCNT_MAGIC | priv->count, priv->base + WTCNT);
+	writew(WTCSR_MAGIC | WTSCR_WT | WTSCR_TME |
+	       WTSCR_CKS(priv->cks), priv->base + WTCSR);
 
 	return 0;
 }
@@ -75,7 +103,12 @@  static int rza_wdt_ping(struct watchdog_device *wdev)
 {
 	struct rza_wdt *priv = watchdog_get_drvdata(wdev);
 
-	writew(WTCNT_MAGIC | 0, priv->base + WTCNT);
+	if (priv->timeout != wdev->timeout)
+		rza_wdt_calc_timeout(priv, wdev->timeout);
+
+	writew(WTCNT_MAGIC | priv->count, priv->base + WTCNT);
+
+	pr_debug("%s: timeout = %d\n", __func__, wdev->timeout);
 
 	return 0;
 }
@@ -130,6 +163,7 @@  static int rza_wdt_probe(struct platform_device *pdev)
 	struct resource *res;
 	unsigned long rate;
 	int ret;
+	struct device_node *np = pdev->dev.of_node;
 
 	priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL);
 	if (!priv)
@@ -150,20 +184,29 @@  static int rza_wdt_probe(struct platform_device *pdev)
 		return -ENOENT;
 	}
 
-	/* Assume slowest clock rate possible (CKS=7) */
-	rate /= 16384;
+	if (of_device_is_compatible(np, "renesas,r7s9210-wdt"))
+		priv->ext_cks = true;
 
 	priv->wdev.info = &rza_wdt_ident,
 	priv->wdev.ops = &rza_wdt_ops,
 	priv->wdev.parent = &pdev->dev;
 
-	/*
-	 * Since the max possible timeout of our 8-bit count register is less
-	 * than a second, we must use max_hw_heartbeat_ms.
-	 */
-	priv->wdev.max_hw_heartbeat_ms = (1000 * U8_MAX) / rate;
-	dev_dbg(&pdev->dev, "max hw timeout of %dms\n",
-		 priv->wdev.max_hw_heartbeat_ms);
+	if (priv->ext_cks) {
+		/* Assume slowest clock rate possible (CKS=0xF) */
+		priv->wdev.max_timeout = (4194304 * U8_MAX) / rate;
+	} else {
+		/* Assume slowest clock rate possible (CKS=7) */
+		rate /= 16384;
+
+		/*
+		 * Since the max possible timeout of our 8-bit count
+		 * register is less than a second, we must use
+		 * max_hw_heartbeat_ms.
+		 */
+		priv->wdev.max_hw_heartbeat_ms = (1000 * U8_MAX) / rate;
+		dev_dbg(&pdev->dev, "max hw timeout of %dms\n",
+			 priv->wdev.max_hw_heartbeat_ms);
+	}
 
 	priv->wdev.min_timeout = 1;
 	priv->wdev.timeout = DEFAULT_TIMEOUT;