fxls8962af_fifo_flush() uses indio_dev->active_scan_mask (with iio_for_each_active_channel()) without making sure the indio_dev stays in buffer mode. There is a race if indio_dev exits buffer mode in the middle of the interrupt that flushes the fifo. Fix this by calling synchronize_irq() to ensure that no interrupt is currently running when disabling buffer mode.
Unable to handle kernel NULL pointer dereference at virtual address 00000000 when read [...] _find_first_bit_le from fxls8962af_fifo_flush+0x17c/0x290 fxls8962af_fifo_flush from fxls8962af_interrupt+0x80/0x178 fxls8962af_interrupt from irq_thread_fn+0x1c/0x7c irq_thread_fn from irq_thread+0x110/0x1f4 irq_thread from kthread+0xe0/0xfc kthread from ret_from_fork+0x14/0x2c
Fixes: 79e3a5bdd9ef ("iio: accel: fxls8962af: add hw buffered sampling") Cc: stable@vger.kernel.org Suggested-by: David Lechner dlechner@baylibre.com Signed-off-by: Sean Nyekjaer sean@geanix.com --- Changes in v2: - As per David's suggestion; switched to use synchronize_irq() instead. - Link to v1: https://lore.kernel.org/r/20250524-fxlsrace-v1-1-dec506dc87ae@geanix.com --- drivers/iio/accel/fxls8962af-core.c | 2 ++ 1 file changed, 2 insertions(+)
diff --git a/drivers/iio/accel/fxls8962af-core.c b/drivers/iio/accel/fxls8962af-core.c index 6d23da3e7aa22c61f2d9348bb91d70cc5719a732..f2558fba491dffa78b26d47d2cd9f1f4d9811f54 100644 --- a/drivers/iio/accel/fxls8962af-core.c +++ b/drivers/iio/accel/fxls8962af-core.c @@ -866,6 +866,8 @@ static int fxls8962af_buffer_predisable(struct iio_dev *indio_dev) if (ret) return ret;
+ synchronize_irq(data->irq); + ret = __fxls8962af_fifo_set_mode(data, false);
if (data->enable_event)
--- base-commit: 5c3fcb36c92443a9a037683626a2e43d8825f783 change-id: 20250524-fxlsrace-f4d20e29fb29
Best regards,
On 6/3/25 7:25 AM, Sean Nyekjaer wrote:
fxls8962af_fifo_flush() uses indio_dev->active_scan_mask (with iio_for_each_active_channel()) without making sure the indio_dev stays in buffer mode. There is a race if indio_dev exits buffer mode in the middle of the interrupt that flushes the fifo. Fix this by calling synchronize_irq() to ensure that no interrupt is currently running when disabling buffer mode.
Unable to handle kernel NULL pointer dereference at virtual address 00000000 when read [...] _find_first_bit_le from fxls8962af_fifo_flush+0x17c/0x290 fxls8962af_fifo_flush from fxls8962af_interrupt+0x80/0x178 fxls8962af_interrupt from irq_thread_fn+0x1c/0x7c irq_thread_fn from irq_thread+0x110/0x1f4 irq_thread from kthread+0xe0/0xfc kthread from ret_from_fork+0x14/0x2c
Fixes: 79e3a5bdd9ef ("iio: accel: fxls8962af: add hw buffered sampling") Cc: stable@vger.kernel.org Suggested-by: David Lechner dlechner@baylibre.com Signed-off-by: Sean Nyekjaer sean@geanix.com
Changes in v2:
- As per David's suggestion; switched to use synchronize_irq() instead.
- Link to v1: https://lore.kernel.org/r/20250524-fxlsrace-v1-1-dec506dc87ae@geanix.com
Were you able to find a way to reproduce the bug well enough to test this?
On Tue, Jun 03, 2025 at 08:29:51AM +0100, David Lechner wrote:
On 6/3/25 7:25 AM, Sean Nyekjaer wrote:
fxls8962af_fifo_flush() uses indio_dev->active_scan_mask (with iio_for_each_active_channel()) without making sure the indio_dev stays in buffer mode. There is a race if indio_dev exits buffer mode in the middle of the interrupt that flushes the fifo. Fix this by calling synchronize_irq() to ensure that no interrupt is currently running when disabling buffer mode.
Unable to handle kernel NULL pointer dereference at virtual address 00000000 when read [...] _find_first_bit_le from fxls8962af_fifo_flush+0x17c/0x290 fxls8962af_fifo_flush from fxls8962af_interrupt+0x80/0x178 fxls8962af_interrupt from irq_thread_fn+0x1c/0x7c irq_thread_fn from irq_thread+0x110/0x1f4 irq_thread from kthread+0xe0/0xfc kthread from ret_from_fork+0x14/0x2c
Fixes: 79e3a5bdd9ef ("iio: accel: fxls8962af: add hw buffered sampling") Cc: stable@vger.kernel.org Suggested-by: David Lechner dlechner@baylibre.com Signed-off-by: Sean Nyekjaer sean@geanix.com
Changes in v2:
- As per David's suggestion; switched to use synchronize_irq() instead.
- Link to v1: https://lore.kernel.org/r/20250524-fxlsrace-v1-1-dec506dc87ae@geanix.com
Were you able to find a way to reproduce the bug well enough to test this?
Yeah. Sorry I didn't included the reproducer :)
I added: diff --git a/drivers/iio/accel/fxls8962af-core.c b/drivers/iio/accel/fxls8962af-core.c index f2558fba491d..ce9a14245f83 100644 --- a/drivers/iio/accel/fxls8962af-core.c +++ b/drivers/iio/accel/fxls8962af-core.c @@ -1040,6 +1040,8 @@ static irqreturn_t fxls8962af_interrupt(int irq, void *p) unsigned int reg; int ret;
+ usleep_range(10000, 15000); + ret = regmap_read(data->regmap, FXLS8962AF_INT_STATUS, ®); if (ret) return IRQ_NONE;
And it was reproduceable within 10 secs: root@localhost:/sys/bus/iio/devices/iio:device0/buffer0# while true; do echo 1 > enable; sleep 1; echo 0 > enable; sleep 1; done
With synchronize_irq(data->irq); I have not been able to reproduce it.
/Sean
linux-stable-mirror@lists.linaro.org