Shutdown bus function might be called on the unbound device, so add a
check if there is a driver before calling its shutdown function.
This fixes following kernel panic obserbed during system reboot:
Unable to handle kernel NULL pointer dereference at virtual address
0000000000000018
...
Call trace:
spi_shutdown+0x10/0x38
kernel_restart_prepare+0x34/0x40
kernel_restart+0x14/0x88
__do_sys_reboot+0x148/0x248
__arm64_sys_reboot+0x1c/0x28
el0_svc_common.constprop.3+0x74/0x198
do_el0_svc+0x20/0x98
el0_sync_handler+0x140/0x1a8
el0_sync+0x140/0x180
Code:
f9403402 d1008041 f100005f 9a9f1021 (
f9400c21)
---[ end trace
266c07205a2d632e ]---
Fixes: 9db34ee64ce4 (spi: Use bus_type functions for probe, remove and shutdown)
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Link: https://lore.kernel.org/r/20201124131523.32287-1-m.szyprowski@samsung.com
Signed-off-by: Mark Brown <broonie@kernel.org>
static void spi_shutdown(struct device *dev)
{
- const struct spi_driver *sdrv = to_spi_driver(dev->driver);
+ if (dev->driver) {
+ const struct spi_driver *sdrv = to_spi_driver(dev->driver);
- if (sdrv->shutdown)
- sdrv->shutdown(to_spi_device(dev));
+ if (sdrv->shutdown)
+ sdrv->shutdown(to_spi_device(dev));
+ }
}
struct bus_type spi_bus_type = {