[PATCH 1/2] pinctrl: sunxi: Forward calls to irq_set_irq_wake

[Samuel Holland]

The pinctrl irqchip may be connected to an irqchip that implements the
.irq_set_wake callback, such as the R_INTC on A31 and newer sunxi SoCs.
In order for GPIOs to be able to trigger wakeup, the IRQ from the
pinctrl to the upper irqchip must also be enabled for wakeup. Since the
kernel's IRQ core already manages the "wake_depth" of each IRQ, no
additional accounting is needed in the pinctrl driver.

Signed-off-by: Samuel Holland <sam...@sholland.org>
---
drivers/pinctrl/sunxi/pinctrl-sunxi.c | 14 ++++++++++++--
1 file changed, 12 insertions(+), 2 deletions(-)

diff --git a/drivers/pinctrl/sunxi/pinctrl-sunxi.c b/drivers/pinctrl/sunxi/pinctrl-sunxi.c
index 0cbca30b75dc..df79da76321e 100644
--- a/drivers/pinctrl/sunxi/pinctrl-sunxi.c
+++ b/drivers/pinctrl/sunxi/pinctrl-sunxi.c
@@ -13,6 +13,7 @@
#include <linux/io.h>
#include <linux/clk.h>
#include <linux/gpio/driver.h>
+#include <linux/interrupt.h>
#include <linux/irqdomain.h>
#include <linux/irqchip/chained_irq.h>
#include <linux/export.h>
@@ -1058,6 +1059,14 @@ static void sunxi_pinctrl_irq_ack_unmask(struct irq_data *d)
sunxi_pinctrl_irq_unmask(d);
}

+static int sunxi_pinctrl_irq_set_wake(struct irq_data *d, unsigned int on)
+{
+ struct sunxi_pinctrl *pctl = irq_data_get_irq_chip_data(d);
+ u8 bank = d->hwirq / IRQ_PER_BANK;
+
+ return irq_set_irq_wake(pctl->irq[bank], on);
+}
+
static struct irq_chip sunxi_pinctrl_edge_irq_chip = {
.name = "sunxi_pio_edge",
.irq_ack = sunxi_pinctrl_irq_ack,
@@ -1066,7 +1075,7 @@ static struct irq_chip sunxi_pinctrl_edge_irq_chip = {
.irq_request_resources = sunxi_pinctrl_irq_request_resources,
.irq_release_resources = sunxi_pinctrl_irq_release_resources,
.irq_set_type = sunxi_pinctrl_irq_set_type,
- .flags = IRQCHIP_SKIP_SET_WAKE,
+ .irq_set_wake = sunxi_pinctrl_irq_set_wake,
};

static struct irq_chip sunxi_pinctrl_level_irq_chip = {
@@ -1081,7 +1090,8 @@ static struct irq_chip sunxi_pinctrl_level_irq_chip = {
.irq_request_resources = sunxi_pinctrl_irq_request_resources,
.irq_release_resources = sunxi_pinctrl_irq_release_resources,
.irq_set_type = sunxi_pinctrl_irq_set_type,
- .flags = IRQCHIP_SKIP_SET_WAKE | IRQCHIP_EOI_THREADED |
+ .irq_set_wake = sunxi_pinctrl_irq_set_wake,
+ .flags = IRQCHIP_EOI_THREADED |
IRQCHIP_EOI_IF_HANDLED,
};

--
2.23.0

[Samuel Holland]

All,

On 1/17/20 3:33 PM, Samuel Holland wrote:
> The pinctrl irqchip may be connected to an irqchip that implements the
> .irq_set_wake callback, such as the R_INTC on A31 and newer sunxi SoCs.
> In order for GPIOs to be able to trigger wakeup, the IRQ from the
> pinctrl to the upper irqchip must also be enabled for wakeup. Since the
> kernel's IRQ core already manages the "wake_depth" of each IRQ, no
> additional accounting is needed in the pinctrl driver.

Ping. Any comments on these two patches?

Thanks,
Samuel Holland

[Linus Walleij]

On Fri, Jan 17, 2020 at 10:33 PM Samuel Holland <sam...@sholland.org> wrote:

> The pinctrl irqchip may be connected to an irqchip that implements the
> .irq_set_wake callback, such as the R_INTC on A31 and newer sunxi SoCs.
> In order for GPIOs to be able to trigger wakeup, the IRQ from the
> pinctrl to the upper irqchip must also be enabled for wakeup. Since the
> kernel's IRQ core already manages the "wake_depth" of each IRQ, no
> additional accounting is needed in the pinctrl driver.
>
> Signed-off-by: Samuel Holland <sam...@sholland.org>

Patch applied with Maxime's ACK.

Yours,
Linus Walleij