Re: [PATCH] mm/gup: Let __get_user_pages_locked() return -EINTR for fatal signal

[Peter Zijlstra]

On Wed, Apr 08, 2020 at 11:59:24AM -0400, Peter Xu wrote:
> From: Hillf Danton <hdanton@xxxxxxxx>
> 
> __get_user_pages_locked() will return 0 instead of -EINTR after commit
> 4426e945df588 which added extra code to allow gup detect fatal signal
> faster.  Restore that behavior.
> 
> Cc: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
> Fixes: 4426e945df58 ("mm/gup: allow VM_FAULT_RETRY for multiple times")
> Reported-by: syzbot+3be1a33f04dc782e9fd5@xxxxxxxxxxxxxxxxxxxxxxxxx
> Signed-off-by: Hillf Danton <hdanton@xxxxxxxx>
> Acked-by: Michal Hocko <mhocko@xxxxxxxx>
> Signed-off-by: Peter Xu <peterx@xxxxxxxxxx>
> ---
> 
> PS. Patch verified with syzbot.
> 
> Signed-off-by: Peter Xu <peterx@xxxxxxxxxx>
> ---
>  mm/gup.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/gup.c b/mm/gup.c
> index afce0bc47e70..6076df8e04a4 100644
> --- a/mm/gup.c
> +++ b/mm/gup.c
> @@ -1326,8 +1326,11 @@ static __always_inline long __get_user_pages_locked(struct task_struct *tsk,
>  		 * start trying again otherwise it can loop forever.
>  		 */
>  
> -		if (fatal_signal_pending(current))
> +		if (fatal_signal_pending(current)) {
> +			if (!pages_done)
> +				pages_done = -EINTR;

Why -EINTR here and -ERESTARTSYS at the other site?

>  			break;
> +		}
>  
>  		ret = down_read_killable(&mm->mmap_sem);
>  		if (ret) {
> -- 
> 2.24.1
>