Re: [patch] mm: update alloc_flags after oom killer has been called

[KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>]

> It is possible for the oom killer to select current as the task to kill.  
> When this happens, alloc_flags needs to be updated accordingly to set 
> ALLOC_NO_WATERMARKS so the subsequent allocation attempt may use memory 
> reserves as the result of its thread having TIF_MEMDIE set if the 
> allocation is not __GFP_NOMEMALLOC.
> 
> Acked-by: Mel Gorman <mel@csn.ul.ie>
> Signed-off-by: David Rientjes <rientjes@google.com>
> ---
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -1756,6 +1756,7 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
>  
>  	wake_all_kswapd(order, zonelist, high_zoneidx);
>  
> +restart:
>  	/*
>  	 * OK, we're below the kswapd watermark and have kicked background
>  	 * reclaim. Now things get more complex, so set up alloc_flags according
> @@ -1763,7 +1764,6 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
>  	 */
>  	alloc_flags = gfp_to_alloc_flags(gfp_mask);
>  
> -restart:
>  	/* This is the last chance, in general, before the goto nopage. */
>  	page = get_page_from_freelist(gfp_mask, nodemask, order, zonelist,
>  			high_zoneidx, alloc_flags & ~ALLOC_NO_WATERMARKS,

Looks good. thanks.
	Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>