Re: [PATCH] mm: remove __GFP_NOFAIL is deprecated comment

[Michal Hocko <mhocko@kernel.org>]

On Thu 25-02-16 13:36:11, Nikolay Borisov wrote:
[...]
> > +		/*
> > +		 * We most definitely don't want callers attempting to
> > +		 * allocate greater than order-1 page units with __GFP_NOFAIL.
> > +		 */
> > +		WARN_ON_ONCE(unlikely(gfp_flags & __GFP_NOFAIL) && (order > 1));
> 
> WARN_ON_ONCE already includes an unlikely in its definition:
> http://lxr.free-electrons.com/source/include/asm-generic/bug.h#L109

OK, I just wanted to keep the condition untouched but you are right the
unlikely can be removed.

> Reviewed-by: Nikolay Borisov <kernel@kyup.com>

Thanks!
---
>From 6ddc73c6d659224e996d1060974f5a14a1c5bf7d Mon Sep 17 00:00:00 2001
From: Michal Hocko <mhocko@suse.com>
Date: Thu, 25 Feb 2016 11:29:12 +0100
Subject: [PATCH] mm: remove __GFP_NOFAIL is deprecated comment

647757197cd3 ("mm: clarify __GFP_NOFAIL deprecation status") was
incomplete and didn't remove the comment about __GFP_NOFAIL being
deprecated in buffered_rmqueue. Let's get rid of this leftover
but keep the WARN_ON_ONCE for order > 1 because we should really
discourage from using __GFP_NOFAIL with higher order allocations
because those are just too subtle.

Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Nikolay Borisov <kernel@kyup.com>
---
 mm/page_alloc.c | 18 +++++-------------
 1 file changed, 5 insertions(+), 13 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 1993894b4219..b5b7a7b0cad0 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -2347,19 +2347,11 @@ struct page *buffered_rmqueue(struct zone *preferred_zone,
 		list_del(&page->lru);
 		pcp->count--;
 	} else {
-		if (unlikely(gfp_flags & __GFP_NOFAIL)) {
-			/*
-			 * __GFP_NOFAIL is not to be used in new code.
-			 *
-			 * All __GFP_NOFAIL callers should be fixed so that they
-			 * properly detect and handle allocation failures.
-			 *
-			 * We most definitely don't want callers attempting to
-			 * allocate greater than order-1 page units with
-			 * __GFP_NOFAIL.
-			 */
-			WARN_ON_ONCE(order > 1);
-		}
+		/*
+		 * We most definitely don't want callers attempting to
+		 * allocate greater than order-1 page units with __GFP_NOFAIL.
+		 */
+		WARN_ON_ONCE((gfp_flags & __GFP_NOFAIL) && (order > 1));
 		spin_lock_irqsave(&zone->lock, flags);
 
 		page = NULL;
-- 
2.7.0


-- 
Michal Hocko
SUSE Labs