Re: [PATCH 5/6] mm/page_alloc: Give GFP_ATOMIC and non-blocking allocations access to reserves

[Mel Gorman <mgorman@techsingularity.net>]

On Thu, Dec 08, 2022 at 07:07:06PM +0100, Vlastimil Babka wrote:
> On 11/29/22 16:17, Mel Gorman wrote:
> > @@ -4846,28 +4846,30 @@ gfp_to_alloc_flags(gfp_t gfp_mask, unsigned int order)
> >  	 * The caller may dip into page reserves a bit more if the caller
> >  	 * cannot run direct reclaim, or if the caller has realtime scheduling
> >  	 * policy or is asking for __GFP_HIGH memory.  GFP_ATOMIC requests will
> > -	 * set both ALLOC_HARDER (__GFP_ATOMIC) and ALLOC_MIN_RESERVE(__GFP_HIGH).
> > +	 * set both ALLOC_NON_BLOCK and ALLOC_MIN_RESERVE(__GFP_HIGH).
> >  	 */
> >  	alloc_flags |= (__force int)
> >  		(gfp_mask & (__GFP_HIGH | __GFP_KSWAPD_RECLAIM));
> >  
> > -	if (gfp_mask & __GFP_ATOMIC) {
> > +	if (!(gfp_mask & __GFP_RECLAIM)) {
> 
> This is supposed to be __GFP_DIRECT_RECLAIM right? Otherwise that includes
> also __GFP_KSWAPD_RECLAIM and GFP_ATOMIC sets that one...
> 

Yes

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 0b2093d17b48..2217bab2dbb2 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -4856,7 +4856,7 @@ gfp_to_alloc_flags(gfp_t gfp_mask, unsigned int order)
 	alloc_flags |= (__force int)
 		(gfp_mask & (__GFP_HIGH | __GFP_KSWAPD_RECLAIM));
 
-	if (!(gfp_mask & __GFP_RECLAIM)) {
+	if (!(gfp_mask & __GFP_DIRECT_RECLAIM)) {
 		/*
 		 * Not worth trying to allocate harder for __GFP_NOMEMALLOC even
 		 * if it can't schedule.

-- 
Mel Gorman
SUSE Labs