[PATCH v2 0/2] Input: gpio-keys - Fixes for interrupt callbacks

[PATCH v2 0/2] Input: gpio-keys - Fixes for interrupt callbacks

[Gatien Chevallier]

This patchset fixes a "BUG: sleeping function called from invalid
context" when CFG_PREEMPT_RT is enabled and a possible concurrent access
on button data between gpio_keys_irq_timer() and gpio_keys_irq_isr().

Signed-off-by: Gatien Chevallier <gatien.chevallier@foss.st.com>
---
Changes in v2:
- Split fixes into two commits
- Use guard() macro to handle spinlock
- Link to v1: https://lore.kernel.org/r/20250526-gpio_keys_preempt_rt-v1-1-09ddadf8e19d@foss.st.com

---
Fabrice Gasnier (1):
      Input: gpio-keys - fix a sleep while atomic with PREEMPT_RT

Gatien Chevallier (1):
      Input: gpio-keys - fix possible concurrent access in gpio_keys_irq_timer()

 drivers/input/keyboard/gpio_keys.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
---
base-commit: 0ff41df1cb268fc69e703a08a57ee14ae967d0ca
change-id: 20250526-gpio_keys_preempt_rt-10619c8fa916

Best regards,
-- 
Gatien Chevallier <gatien.chevallier@foss.st.com>

[PATCH v2 1/2] Input: gpio-keys - fix a sleep while atomic with PREEMPT_RT

[Gatien Chevallier]

From: Fabrice Gasnier <fabrice.gasnier@foss.st.com>

When enabling PREEMPT_RT, the gpio_keys_irq_timer() callback runs in
hard irq context, but the input_event() takes a spin_lock, which isn't
allowed there as it is converted to a rt_spin_lock().

[ 4054.289999] BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48
[ 4054.290028] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/0
...
[ 4054.290195]  __might_resched+0x13c/0x1f4
[ 4054.290209]  rt_spin_lock+0x54/0x11c
[ 4054.290219]  input_event+0x48/0x80
[ 4054.290230]  gpio_keys_irq_timer+0x4c/0x78
[ 4054.290243]  __hrtimer_run_queues+0x1a4/0x438
[ 4054.290257]  hrtimer_interrupt+0xe4/0x240
[ 4054.290269]  arch_timer_handler_phys+0x2c/0x44
[ 4054.290283]  handle_percpu_devid_irq+0x8c/0x14c
[ 4054.290297]  handle_irq_desc+0x40/0x58
[ 4054.290307]  generic_handle_domain_irq+0x1c/0x28
[ 4054.290316]  gic_handle_irq+0x44/0xcc

Considering the gpio_keys_irq_isr() can run in any context, e.g. it can
be threaded, it seems there's no point in requesting the timer isr to
run in hard irq context.

Relax the hrtimer not to use the hard context.

Fixes: 019002f20cb5 ("Input: gpio-keys - use hrtimer for release timer")
Suggested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
Signed-off-by: Gatien Chevallier <gatien.chevallier@foss.st.com>
---
 drivers/input/keyboard/gpio_keys.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/input/keyboard/gpio_keys.c b/drivers/input/keyboard/gpio_keys.c
index 5c39a217b94c8ad03a8542380eed741fccdca5da..d884538107c9654d21e37fa20e0b3b6cd9a25ee4 100644
--- a/drivers/input/keyboard/gpio_keys.c
+++ b/drivers/input/keyboard/gpio_keys.c
@@ -486,7 +486,7 @@ static irqreturn_t gpio_keys_irq_isr(int irq, void *dev_id)
 	if (bdata->release_delay)
 		hrtimer_start(&bdata->release_timer,
 			      ms_to_ktime(bdata->release_delay),
-			      HRTIMER_MODE_REL_HARD);
+			      HRTIMER_MODE_REL);
 out:
 	return IRQ_HANDLED;
 }
@@ -628,7 +628,7 @@ static int gpio_keys_setup_key(struct platform_device *pdev,
 
 		bdata->release_delay = button->debounce_interval;
 		hrtimer_setup(&bdata->release_timer, gpio_keys_irq_timer,
-			      CLOCK_REALTIME, HRTIMER_MODE_REL_HARD);
+			      CLOCK_REALTIME, HRTIMER_MODE_REL);
 
 		isr = gpio_keys_irq_isr;
 		irqflags = 0;

-- 
2.25.1

[PATCH v2 2/2] Input: gpio-keys - fix possible concurrent access in gpio_keys_irq_timer()

[Gatien Chevallier]

gpio_keys_irq_isr() and gpio_keys_irq_timer() access the same resources.
There could be a concurrent access if a GPIO interrupt occurs in parallel
of a HR timer interrupt.

Guard back those resources with a spinlock.

Fixes: 019002f20cb5 ("Input: gpio-keys - use hrtimer for release timer")
Signed-off-by: Gatien Chevallier <gatien.chevallier@foss.st.com>
---
 drivers/input/keyboard/gpio_keys.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/input/keyboard/gpio_keys.c b/drivers/input/keyboard/gpio_keys.c
index d884538107c9654d21e37fa20e0b3b6cd9a25ee4..f9db86da0818b24d3b6611b4d4e23c96846dbf18 100644
--- a/drivers/input/keyboard/gpio_keys.c
+++ b/drivers/input/keyboard/gpio_keys.c
@@ -449,6 +449,8 @@ static enum hrtimer_restart gpio_keys_irq_timer(struct hrtimer *t)
 						      release_timer);
 	struct input_dev *input = bdata->input;
 
+	guard(spinlock_irqsave)(&bdata->lock);
+
 	if (bdata->key_pressed) {
 		input_report_key(input, *bdata->code, 0);
 		input_sync(input);

-- 
2.25.1

Re: [PATCH v2 0/2] Input: gpio-keys - Fixes for interrupt callbacks

[Dmitry Torokhov]

On Wed, May 28, 2025 at 10:50:20AM +0200, Gatien Chevallier wrote:
> This patchset fixes a "BUG: sleeping function called from invalid
> context" when CFG_PREEMPT_RT is enabled and a possible concurrent access
> on button data between gpio_keys_irq_timer() and gpio_keys_irq_isr().
> 
> Signed-off-by: Gatien Chevallier <gatien.chevallier@foss.st.com>

Applied, thank you.

-- 
Dmitry