Re: [PATCH 4.19 108/125] USB: yurex: Fix bad gfp argument

[Alan Stern]

On Wed, Sep 02, 2020 at 02:58:27PM +0200, Pavel Machek wrote:
> Hi!
> 
> > The syzbot fuzzer identified a bug in the yurex driver: It passes
> > GFP_KERNEL as a memory-allocation flag to usb_submit_urb() at a time
> > when its state is TASK_INTERRUPTIBLE, not TASK_RUNNING:
> 
> Yeah, and instead of fixing the bug, patch papers over it, reducing
> reliability of the driver in the process.
> 
> > This patch changes the call to use GFP_ATOMIC instead of GFP_KERNEL.
> 
> Fixing it properly should be as simple as moving prepare_to_wait down,
> no?

Sigh.  That never occurred to me, but of course it is right.

Acked-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>

Alan Stern

> Signed-off-by: Pavel Machek (CIP) <pavel@xxxxxxx>
> 
> diff --git a/drivers/usb/misc/yurex.c b/drivers/usb/misc/yurex.c
> index 785080f79073..5fbbb57e6e95 100644
> --- a/drivers/usb/misc/yurex.c
> +++ b/drivers/usb/misc/yurex.c
> @@ -489,10 +489,10 @@ static ssize_t yurex_write(struct file *file, const char __user *user_buffer,
>  	}
>  
>  	/* send the data as the control msg */
> -	prepare_to_wait(&dev->waitq, &wait, TASK_INTERRUPTIBLE);
>  	dev_dbg(&dev->interface->dev, "%s - submit %c\n", __func__,
>  		dev->cntl_buffer[0]);
> -	retval = usb_submit_urb(dev->cntl_urb, GFP_ATOMIC);
> +	retval = usb_submit_urb(dev->cntl_urb, GFP_KERNEL);
> +	prepare_to_wait(&dev->waitq, &wait, TASK_INTERRUPTIBLE);
>  	if (retval >= 0)
>  		timeout = schedule_timeout(YUREX_WRITE_TIMEOUT);
>  	finish_wait(&dev->waitq, &wait);
> 
> 
> Best regards,
> 									Pavel
> 
> -- 
> (english) http://www.livejournal.com/~pavelmachek
> (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html