Re: [PATCH] tty/n_hdlc: fix __might_sleep warning

[Tetsuo Handa]

Hello, Greg.

We attempted a cleanup but we failed. Thus, will you apply this patch for now?

On 2019/01/02 5:28, Paul Fulghum wrote:
> Fix __might_sleep warning in tty/n_hdlc.c read due to copy_to_user call while current is TASK_INTERRUPTIBLE.
> This is a false positive since the code path does not depend on current state remaining TASK_INTERRUPTIBLE.
> The loop breaks out and sets TASK_RUNNING after calling copy_to_user. 
> This patch supresses the warning by setting TASK_RUNNING before calling copy_to_user.
> 
> [1] https://syzkaller.appspot.com/bug?id=17d5de7f1fcab794cb8c40032f893f52de899324
> 
> Signed-off-by: Paul Fulghum <paulkf@xxxxxxxxxxxxx>
> Reported-by: syzbot <syzbot+c244af085a0159d22879@xxxxxxxxxxxxxxxxxxxxxxxxx>
> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Cc: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
> Cc: Arnd Bergmann <arnd@xxxxxxxx>
> Cc: Alan Cox <alan@xxxxxxxxxxxxxxxxxxx>
> â
> --- a/drivers/tty/n_hdlc.c	2018-12-23 15:55:59.000000000 -0800
> +++ b/drivers/tty/n_hdlc.c	2019-01-01 11:44:47.148153954 -0800
> @@ -597,6 +597,7 @@ static ssize_t n_hdlc_tty_read(struct tt
>  				/* too large for caller's buffer */
>  				ret = -EOVERFLOW;
>  			} else {
> +				__set_current_state(TASK_RUNNING);
>  				if (copy_to_user(buf, rbuf->buf, rbuf->count))
>  					ret = -EFAULT;
>  				else