If a PCA953x gpio was used as an interrupt and then released, the shutdown function was trying to extract the pca953x_chip pointer directly from the irq_data, but in reality was getting the gpio_chip structure. The net effect was that the subsequent writes to the data structure corrupted data in the gpio_chip structure, which wasn't immediately obvious until attempting to use the GPIO again in the future, at which point the kernel panics. This fix correctly extracts the pca953x_chip structure via the gpio_chip structure, as is correctly done in the other irq functions. Fixes: 0a70fe00efea ("gpio: pca953x: Clear irq trigger type on irq shutdown") Cc: stable@xxxxxxxxxxxxxxx Signed-off-by: Mark Walton <mark.walton@xxxxxxxxxxxxx> --- drivers/gpio/gpio-pca953x.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpio/gpio-pca953x.c b/drivers/gpio/gpio-pca953x.c index caf7dd1..6bd55a4 100644 --- a/drivers/gpio/gpio-pca953x.c +++ b/drivers/gpio/gpio-pca953x.c @@ -659,7 +659,8 @@ static int pca953x_irq_set_type(struct irq_data *d, unsigned int type) static void pca953x_irq_shutdown(struct irq_data *d) { - struct pca953x_chip *chip = irq_data_get_irq_chip_data(d); + struct gpio_chip *gc = irq_data_get_irq_chip_data(d); + struct pca953x_chip *chip = gpiochip_get_data(gc); u8 mask = 1 << (d->hwirq % BANK_SZ); chip->irq_trig_raise[d->hwirq / BANK_SZ] &= ~mask; -- 2.7.4