linux-kernel - Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of the freelist in unset_migratetype

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 25 Sep 2020 10:39:33 +0200
From:   Vlastimil Babka <vbabka@...e.cz>
To:     David Hildenbrand <david@...hat.com>,
        Wei Yang <richard.weiyang@...ux.alibaba.com>
Cc:     linux-kernel@...r.kernel.org, linux-mm@...ck.org,
        linux-hyperv@...r.kernel.org, xen-devel@...ts.xenproject.org,
        linux-acpi@...r.kernel.org,
        Andrew Morton <akpm@...ux-foundation.org>,
        Alexander Duyck <alexander.h.duyck@...ux.intel.com>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Michal Hocko <mhocko@...nel.org>,
        Dave Hansen <dave.hansen@...el.com>,
        Oscar Salvador <osalvador@...e.de>,
        Mike Rapoport <rppt@...nel.org>,
        Scott Cheloha <cheloha@...ux.ibm.com>,
        Michael Ellerman <mpe@...erman.id.au>
Subject: Re: [PATCH RFC 3/4] mm/page_alloc: always move pages to the tail of
 the freelist in unset_migratetype_isolate()

On 9/25/20 10:05 AM, David Hildenbrand wrote:
>>>>  static inline void del_page_from_free_list(struct page *page, struct zone *zone,
>>>>  					   unsigned int order)
>>>>  {
>>>> @@ -2323,7 +2332,7 @@ static inline struct page *__rmqueue_cma_fallback(struct zone *zone,
>>>>   */
>>>>  static int move_freepages(struct zone *zone,
>>>>  			  struct page *start_page, struct page *end_page,
>>>> -			  int migratetype, int *num_movable)
>>>> +			  int migratetype, int *num_movable, bool to_tail)
>>>>  {
>>>>  	struct page *page;
>>>>  	unsigned int order;
>>>> @@ -2354,7 +2363,10 @@ static int move_freepages(struct zone *zone,
>>>>  		VM_BUG_ON_PAGE(page_zone(page) != zone, page);
>>>>  
>>>>  		order = page_order(page);
>>>> -		move_to_free_list(page, zone, order, migratetype);
>>>> +		if (to_tail)
>>>> +			move_to_free_list_tail(page, zone, order, migratetype);
>>>> +		else
>>>> +			move_to_free_list(page, zone, order, migratetype);
>>>>  		page += 1 << order;
>>>>  		pages_moved += 1 << order;
>>>>  	}
>>>> @@ -2362,8 +2374,9 @@ static int move_freepages(struct zone *zone,
>>>>  	return pages_moved;
>>>>  }
>>>>  
>>>> -int move_freepages_block(struct zone *zone, struct page *page,
>>>> -				int migratetype, int *num_movable)
>>>> +static int __move_freepages_block(struct zone *zone, struct page *page,
>>>> +				  int migratetype, int *num_movable,
>>>> +				  bool to_tail)
>>>>  {
>>>>  	unsigned long start_pfn, end_pfn;
>>>>  	struct page *start_page, *end_page;
>>>> @@ -2384,7 +2397,20 @@ int move_freepages_block(struct zone *zone, struct page *page,
>>>>  		return 0;
>>>>  
>>>>  	return move_freepages(zone, start_page, end_page, migratetype,
>>>> -								num_movable);
>>>> +			      num_movable, to_tail);
>>>> +}
>>>> +
>>>> +int move_freepages_block(struct zone *zone, struct page *page,
>>>> +			 int migratetype, int *num_movable)
>>>> +{
>>>> +	return __move_freepages_block(zone, page, migratetype, num_movable,
>>>> +				      false);
>>>> +}
>>>> +
>>>> +int move_freepages_block_tail(struct zone *zone, struct page *page,
>>>> +			      int migratetype)
>>>> +{
>>>> +	return __move_freepages_block(zone, page, migratetype, NULL, true);
>>>>  }
>>>
>>> Likewise, just 5 callers of move_freepages_block(), all in the files you're
>>> already changing, so no need for this wrappers IMHO.
> 
> As long as we don't want to move the implementation to the header, we'll
> need it for the constant propagation to work at compile time (we don't
> really have link-time optimizations). Or am I missing something?

I guess move_freepages_block() is not exactly fast path, so we could do without it.

> Thanks!
>