[PATCH] When ones changes the state of any input pins of a PCA9555 chip before setting up the IRQ mask through i.e. SysFS e.g. echo "both" > /sys/class/gpio/gpioXYZ/edge the epoll_wait shall not exit on the subsequent change of the GPIO state. The reason behind it is that the IRQ status is not being saved when the IRQ is masked.

[Tomasz Kazimierz Motyl <tomasz.motyl666@xxxxxxxxx>]

---
 drivers/gpio/gpio-pca953x.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/gpio/gpio-pca953x.c b/drivers/gpio/gpio-pca953x.c
index 7e76830b3368..088bef902156 100644
--- a/drivers/gpio/gpio-pca953x.c
+++ b/drivers/gpio/gpio-pca953x.c
@@ -716,13 +716,16 @@ static bool pca953x_irq_pending(struct pca953x_chip *chip, u8 *pending)
 		trigger[i] = (cur_stat[i] ^ old_stat[i]) & chip->irq_mask[i];
 		if (trigger[i])
 			trigger_seen = true;
+
+    /* We want the current status recorded in the chip->irq stat regardless the
+     * chip->irq_mask setting in order to have a change detected when the interrupt
+     * mask gets changed i.e. echo "both" > /sys/class/gpioXYZ/edge */
+    chip->irq_stat[i] = cur_stat[i];
 	}
 
 	if (!trigger_seen)
 		return false;
 
-	memcpy(chip->irq_stat, cur_stat, NBANK(chip));
-
 	for (i = 0; i < NBANK(chip); i++) {
 		pending[i] = (old_stat[i] & chip->irq_trig_fall[i]) |
 			(cur_stat[i] & chip->irq_trig_raise[i]);
-- 
2.17.1