Re: [PATCH v2 2/2] gpiolib: protect the GPIO device against being dropped while in use by user-space
From: Andy Shevchenko
Date: Mon Nov 28 2022 - 14:22:13 EST
On Mon, Nov 28, 2022 at 06:52:14PM +0100, Bartosz Golaszewski wrote:
> From: Bartosz Golaszewski <bartosz.golaszewski@xxxxxxxxxx>
>
> While any of the GPIO cdev syscalls is in progress, the kernel can call
> gpiochip_remove() (for instance, when a USB GPIO expander is disconnected)
> which will set gdev->chip to NULL after which any subsequent access will
> cause a crash.
>
> To avoid that: use an RW-semaphore in which the syscalls take it for
> reading (so that we don't needlessly prohibit the user-space from calling
> syscalls simultaneously) while gpiochip_remove() takes it for writing so
> that it can only happen once all syscalls return.
Missed also Dependency (the previous change).
> Fixes: d7c51b47ac11 ("gpio: userspace ABI for reading/writing GPIO lines")
> Fixes: 3c0d9c635ae2 ("gpiolib: cdev: support GPIO_V2_GET_LINE_IOCTL and GPIO_V2_LINE_GET_VALUES_IOCTL")
> Fixes: aad955842d1c ("gpiolib: cdev: support GPIO_V2_GET_LINEINFO_IOCTL and GPIO_V2_GET_LINEINFO_WATCH_IOCTL")
> Fixes: a54756cb24ea ("gpiolib: cdev: support GPIO_V2_LINE_SET_CONFIG_IOCTL")
> Fixes: 7b8e00d98168 ("gpiolib: cdev: support GPIO_V2_LINE_SET_VALUES_IOCTL")
...
> + down_read(&gdev->sem);
>
> - if (!gdev->chip)
> + if (!gdev->chip) {
> + up_read(&gdev->sem);
> return -ENODEV;
> + }
Wouldn't be easier to wrap existing functions (with their renaming) into a new
ones with semaphore?
You can even start that in the previous patch.
--
With Best Regards,
Andy Shevchenko