pinctrl/amd: poll InterruptEnable bits in enable_irq

djkurtz · linusw · commit 4c1de0414a13 · 2018-03-26T11:09:49.000+02:00
In certain cases interrupt enablement will be delayed relative to when
the InterruptEnable bits are written.  One example of this is when
a GPIO's "debounce" logice is first enabled.  After enabling debounce,
there is a 900 us "warm up" period during which InterruptEnable[0]
(bit 11) will read as 0 despite being written 1.  During this time
InterruptSts will not be updated, nor will interrupts be delivered, even
if the GPIO's interrupt configuration has been written to the register.

To work around this delay, poll the InterruptEnable bits after setting
them to ensure interrupts have truly been enabled in hardware before
returning from the irq_enable handler.

Signed-off-by: Daniel Kurtz &lt;djkurtz@chromium.org&gt;
Signed-off-by: Linus Walleij &lt;linus.walleij@linaro.org&gt;
diff --git a/drivers/pinctrl/pinctrl-amd.c b/drivers/pinctrl/pinctrl-amd.c
@@ -348,12 +348,21 @@ static void amd_gpio_irq_enable(struct irq_data *d)
 	unsigned long flags;
 	struct gpio_chip *gc = irq_data_get_irq_chip_data(d);
 	struct amd_gpio *gpio_dev = gpiochip_get_data(gc);
+	u32 mask = BIT(INTERRUPT_ENABLE_OFF) | BIT(INTERRUPT_MASK_OFF);
 
 	raw_spin_lock_irqsave(&gpio_dev->lock, flags);
 	pin_reg = readl(gpio_dev->base + (d->hwirq)*4);
 	pin_reg |= BIT(INTERRUPT_ENABLE_OFF);
 	pin_reg |= BIT(INTERRUPT_MASK_OFF);
 	writel(pin_reg, gpio_dev->base + (d->hwirq)*4);
+	/*
+	 * When debounce logic is enabled it takes ~900 us before interrupts
+	 * can be enabled.  During this "debounce warm up" period the
+	 * "INTERRUPT_ENABLE" bit will read as 0. Poll the bit here until it
+	 * reads back as 1, signaling that interrupts are now enabled.
+	 */
+	while ((readl(gpio_dev->base + (d->hwirq)*4) & mask) != mask)
+		continue;
 	raw_spin_unlock_irqrestore(&gpio_dev->lock, flags);
 }
 
