linux-kernel - Re: [PATCH] mm/page_alloc: call check_new_pages() while zone spinlock is not held

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]

Message-ID: <3fdd3840-fc97-96a6-8b61-ad1c89b5b403@suse.cz>
Date:   Fri, 4 Mar 2022 09:26:38 +0100
From:   Vlastimil Babka <vbabka@...e.cz>
To:     Eric Dumazet <eric.dumazet@...il.com>,
        Andrew Morton <akpm@...ux-foundation.org>
Cc:     linux-kernel <linux-kernel@...r.kernel.org>,
        linux-mm <linux-mm@...ck.org>,
        Eric Dumazet <edumazet@...gle.com>,
        Mel Gorman <mgorman@...hsingularity.net>,
        Michal Hocko <mhocko@...nel.org>,
        Shakeel Butt <shakeelb@...gle.com>,
        Wei Xu <weixugc@...gle.com>, Greg Thelen <gthelen@...gle.com>,
        Hugh Dickins <hughd@...gle.com>,
        David Rientjes <rientjes@...gle.com>
Subject: Re: [PATCH] mm/page_alloc: call check_new_pages() while zone spinlock
 is not held

On 3/4/22 02:59, Eric Dumazet wrote:
> From: Eric Dumazet <edumazet@...gle.com>
> 
> For high order pages not using pcp, rmqueue() is currently calling
> the costly check_new_pages() while zone spinlock is held.
> 
> This is not needed, we can release the spinlock sooner to reduce
> zone spinlock contention.
> 
> Signed-off-by: Eric Dumazet <edumazet@...gle.com>
> Cc: Mel Gorman <mgorman@...hsingularity.net>
> Cc: Vlastimil Babka <vbabka@...e.cz>
> Cc: Michal Hocko <mhocko@...nel.org>
> Cc: Shakeel Butt <shakeelb@...gle.com>
> Cc: Wei Xu <weixugc@...gle.com>
> Cc: Greg Thelen <gthelen@...gle.com>
> Cc: Hugh Dickins <hughd@...gle.com>
> Cc: David Rientjes <rientjes@...gle.com>
> ---
>  mm/page_alloc.c | 17 +++++++----------
>  1 file changed, 7 insertions(+), 10 deletions(-)

Just a quick reply from what AFAIK immediately see as a issue.

> 
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 3589febc6d31928f850ebe5a4015ddc40e0469f3..0890a65f8cc2259e82bc1f5ba95a592fb30f9fb8 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -3685,7 +3685,6 @@ struct page *rmqueue(struct zone *preferred_zone,
>  			gfp_t gfp_flags, unsigned int alloc_flags,
>  			int migratetype)
>  {
> -	unsigned long flags;
>  	struct page *page;
>  
>  	if (likely(pcp_allowed_order(order))) {
> @@ -3706,10 +3705,12 @@ struct page *rmqueue(struct zone *preferred_zone,
>  	 * 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);
>  
>  	do {
> +		unsigned long flags;
> +
>  		page = NULL;
> +		spin_lock_irqsave(&zone->lock, flags);
>  		/*
>  		 * order-0 request can reach here when the pcplist is skipped
>  		 * due to non-CMA allocation context. HIGHATOMIC area is
> @@ -3723,13 +3724,13 @@ struct page *rmqueue(struct zone *preferred_zone,
>  		}
>  		if (!page)
>  			page = __rmqueue(zone, order, migratetype, alloc_flags);
> -	} while (page && check_new_pages(page, order));
> -	if (!page)
> -		goto failed;
> +		spin_unlock_irqrestore(&zone->lock, flags);
> +		if (!page)
> +			return NULL;
> +	} while (check_new_pages(page, order));
>  
>  	__mod_zone_freepage_state(zone, -(1 << order),
>  				  get_pcppage_migratetype(page));
> -	spin_unlock_irqrestore(&zone->lock, flags);

__mod_zone_freepage_state() outside the lock is AFAIK unsafe.

>  
>  	__count_zid_vm_events(PGALLOC, page_zonenum(page), 1 << order);
>  	zone_statistics(preferred_zone, zone, 1);
> @@ -3743,10 +3744,6 @@ struct page *rmqueue(struct zone *preferred_zone,
>  
>  	VM_BUG_ON_PAGE(page && bad_range(zone, page), page);
>  	return page;
> -
> -failed:
> -	spin_unlock_irqrestore(&zone->lock, flags);
> -	return NULL;
>  }
>  
>  #ifdef CONFIG_FAIL_PAGE_ALLOC