[PATCH AUTOSEL 4.14 3/6] pinctrl: amd: Mask non-wake source pins with interrupt enabled at suspend

[PATCH AUTOSEL 4.14 3/6] pinctrl: amd: Mask non-wake source pins with interrupt enabled at suspend

[Sasha Levin]

From: Mario Limonciello <mario.limonciello@amd.com>

[ Upstream commit 2fff0b5e1a6b9c577b4dd4958902c877159c856b ]

If a pin isn't marked as a wake source processing any interrupts is
just going to destroy battery life.  The APU may wake up from a hardware
sleep state to process the interrupt but not return control to the OS.

Mask interrupt for all non-wake source pins at suspend. They'll be
re-enabled at resume.

Reported-and-tested-by: Marcus Aram <marcus+oss@oxar.nl>
Reported-and-tested-by: Mark Herbert <mark.herbert42@gmail.com>
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2812
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20231203032431.30277-3-mario.limonciello@amd.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/pinctrl/pinctrl-amd.c | 9 +++++++++
 drivers/pinctrl/pinctrl-amd.h | 5 +++++
 2 files changed, 14 insertions(+)

diff --git a/drivers/pinctrl/pinctrl-amd.c b/drivers/pinctrl/pinctrl-amd.c
index 509ba4bceefcb..a5cd88c5dd844 100644
--- a/drivers/pinctrl/pinctrl-amd.c
+++ b/drivers/pinctrl/pinctrl-amd.c
@@ -766,6 +766,15 @@ int amd_gpio_suspend(struct device *dev)
 
 		raw_spin_lock_irqsave(&gpio_dev->lock, flags);
 		gpio_dev->saved_regs[i] = readl(gpio_dev->base + pin * 4) & ~PIN_IRQ_PENDING;
+
+		/* mask any interrupts not intended to be a wake source */
+		if (!(gpio_dev->saved_regs[i] & WAKE_SOURCE)) {
+			writel(gpio_dev->saved_regs[i] & ~BIT(INTERRUPT_MASK_OFF),
+			       gpio_dev->base + pin * 4);
+			pm_pr_dbg("Disabling GPIO #%d interrupt for suspend.\n",
+				  pin);
+		}
+
 		raw_spin_unlock_irqrestore(&gpio_dev->lock, flags);
 	}
 
diff --git a/drivers/pinctrl/pinctrl-amd.h b/drivers/pinctrl/pinctrl-amd.h
index 884f48f7a6a36..065419dabffb5 100644
--- a/drivers/pinctrl/pinctrl-amd.h
+++ b/drivers/pinctrl/pinctrl-amd.h
@@ -80,6 +80,11 @@ struct amd_pingroup {
 	unsigned npins;
 };
 
+#define WAKE_SOURCE	(BIT(WAKE_CNTRL_OFF_S0I3) | \
+			 BIT(WAKE_CNTRL_OFF_S3)   | \
+			 BIT(WAKE_CNTRL_OFF_S4)   | \
+			 BIT(WAKECNTRL_Z_OFF))
+
 struct amd_function {
 	const char *name;
 	const char * const *groups;
-- 
2.43.0

Re: [PATCH AUTOSEL 4.14 3/6] pinctrl: amd: Mask non-wake source pins with interrupt enabled at suspend

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 667 bytes --]

Hi!

> From: Mario Limonciello <mario.limonciello@amd.com>
> 
> [ Upstream commit 2fff0b5e1a6b9c577b4dd4958902c877159c856b ]
> 
> If a pin isn't marked as a wake source processing any interrupts is
> just going to destroy battery life.  The APU may wake up from a hardware
> sleep state to process the interrupt but not return control to the OS.
> 
> Mask interrupt for all non-wake source pins at suspend. They'll be
> re-enabled at resume.

Not sure what is going on here, but it is queued for 6.6 and 4.14, but
not for the kernels between them.

Best regards,
								Pavel
-- 
People of Russia, stop Putin before his war on Ukraine escalates.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: [PATCH AUTOSEL 4.14 3/6] pinctrl: amd: Mask non-wake source pins with interrupt enabled at suspend

[Sasha Levin]

On Tue, Jan 09, 2024 at 12:43:27PM +0100, Pavel Machek wrote:
>Hi!
>
>> From: Mario Limonciello <mario.limonciello@amd.com>
>>
>> [ Upstream commit 2fff0b5e1a6b9c577b4dd4958902c877159c856b ]
>>
>> If a pin isn't marked as a wake source processing any interrupts is
>> just going to destroy battery life.  The APU may wake up from a hardware
>> sleep state to process the interrupt but not return control to the OS.
>>
>> Mask interrupt for all non-wake source pins at suspend. They'll be
>> re-enabled at resume.
>
>Not sure what is going on here, but it is queued for 6.6 and 4.14, but
>not for the kernels between them.

Ack, it should have been dropped from 4.14. Thanks!

-- 
Thanks,
Sasha