2017-12-12 03:59:39

by Jia-Ju Bai

Subject: [BUG] drivers/gpio: some possible sleep-in-atomic bugs

According to drivers/gpio/gpio-grgpio.c, the kernel module may sleep
under a spinlock.
The function call paths are:
grgpio_remove (acquire the spinlock)
gpiochip_remove
gpiochip_sysfs_unregister
device_unregister --> may sleep
mutex_lock --> may sleep

grgpio_remove (acquire the spinlock)
gpiochip_remove
gpiochip_sysfs_unregister
gpiod_free
__gpiod_free
gpiod_unexport
device_unregister --> may sleep
mutex_lock --> may sleep

grgpio_remove (acquire the spinlock)
irq_domain_remove
mutex_lock --> may sleep

grgpio_irq_map (acquire the spinlock)
request_irq --> may sleep

grgpio_irq_unmap (acquire the spinlock)
free_irq --> may sleep


I do not find a good way to fix them, so I only report.
These possible bugs is found by my static analysis tool (DSAC) and
checked by my code review.


Thanks,
Jia-Ju Bai

2017-12-13 16:41:53

by Andreas Larsson

Subject: Re: [BUG] drivers/gpio: some possible sleep-in-atomic bugs

On 2017-12-12 04:59, Jia-Ju Bai wrote:
> According to drivers/gpio/gpio-grgpio.c, the kernel module may sleep
> under a spinlock.
> The function call paths are:
> grgpio_remove (acquire the spinlock)
> gpiochip_remove
> gpiochip_sysfs_unregister
> device_unregister --> may sleep
> mutex_lock --> may sleep
>
> grgpio_remove (acquire the spinlock)
> gpiochip_remove
> gpiochip_sysfs_unregister
> gpiod_free
> __gpiod_free
> gpiod_unexport
> device_unregister --> may sleep
> mutex_lock --> may sleep
>
> grgpio_remove (acquire the spinlock)
> irq_domain_remove
> mutex_lock --> may sleep
>
> grgpio_irq_map (acquire the spinlock)
> request_irq --> may sleep
>
> grgpio_irq_unmap (acquire the spinlock)
> free_irq --> may sleep
>
>
> I do not find a good way to fix them, so I only report.
> These possible bugs is found by my static analysis tool (DSAC) and
> checked by my code review.
Hi!

Thank you for reporting this. I will take a look at it.

--
Best Regards,
Andreas Larsson