Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1319029

Re: [PATCH 1/5] mm: memcontrol: generalize locking for the page->mem_cgroup binding

From Vladimir Davydov <vdavydov@virtuozzo.com>
Newsgroups linux.kernel
Subject Re: [PATCH 1/5] mm: memcontrol: generalize locking for the page->mem_cgroup binding
Date 2016-01-27 15:40 +0100
Message-ID <qVzqi-4RW-33@gated-at.bofh.it> (permalink)
References <qVj2a-1Dd-17@gated-at.bofh.it> <qVj2b-1Dd-31@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


On Tue, Jan 26, 2016 at 04:00:02PM -0500, Johannes Weiner wrote:

> @@ -683,17 +683,17 @@ int __set_page_dirty_buffers(struct page *page)
>  		} while (bh != head);
>  	}
>  	/*
> -	 * Use mem_group_begin_page_stat() to keep PageDirty synchronized with
> -	 * per-memcg dirty page counters.
> +	 * Lock out page->mem_cgroup migration to keep PageDirty
> +	 * synchronized with per-memcg dirty page counters.
>  	 */
> -	memcg = mem_cgroup_begin_page_stat(page);
> +	memcg = lock_page_memcg(page);
>  	newly_dirty = !TestSetPageDirty(page);
>  	spin_unlock(&mapping->private_lock);
>  
>  	if (newly_dirty)
>  		__set_page_dirty(page, mapping, memcg, 1);

Do we really want to pass memcg to __set_page_dirty and then to
account_page_dirtied, increasing stack/regs usage even in case memory
cgroup is disabled? May be, it'd be better to make
mem_cgroup_update_page_stat take a page instead of a memcg?

Thanks,
Vladimir

>  
> -	mem_cgroup_end_page_stat(memcg);
> +	unlock_page_memcg(memcg);
>  
>  	if (newly_dirty)
>  		__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);

Back to linux.kernel | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCH 0/5] mm: workingset: per-cgroup thrash detection Johannes Weiner <hannes@cmpxchg.org> - 2016-01-26 22:10 +0100
  [PATCH 2/5] mm: workingset: #define radix entry eviction mask Johannes Weiner <hannes@cmpxchg.org> - 2016-01-26 22:10 +0100
    Re: [PATCH 2/5] mm: workingset: #define radix entry eviction mask Vladimir Davydov <vdavydov@virtuozzo.com> - 2016-01-27 15:40 +0100
  [PATCH 3/5] mm: workingset: separate shadow unpacking and refault calculation Johannes Weiner <hannes@cmpxchg.org> - 2016-01-26 22:10 +0100
    Re: [PATCH 3/5] mm: workingset: separate shadow unpacking and  refault calculation Vladimir Davydov <vdavydov@virtuozzo.com> - 2016-01-27 15:40 +0100
  [PATCH 4/5] mm: workingset: eviction buckets for bigmem/lowbit machines Johannes Weiner <hannes@cmpxchg.org> - 2016-01-26 22:10 +0100
    Re: [PATCH 4/5] mm: workingset: eviction buckets for bigmem/lowbit  machines Vladimir Davydov <vdavydov@virtuozzo.com> - 2016-01-27 15:50 +0100
  [PATCH 1/5] mm: memcontrol: generalize locking for the page->mem_cgroup binding Johannes Weiner <hannes@cmpxchg.org> - 2016-01-26 22:10 +0100
    Re: [PATCH 1/5] mm: memcontrol: generalize locking for the  page->mem_cgroup binding Vladimir Davydov <vdavydov@virtuozzo.com> - 2016-01-27 15:40 +0100
  [PATCH 5/5] mm: workingset: per-cgroup cache thrash detection Johannes Weiner <hannes@cmpxchg.org> - 2016-01-26 22:10 +0100
    Re: [PATCH 5/5] mm: workingset: per-cgroup cache thrash detection Vladimir Davydov <vdavydov@virtuozzo.com> - 2016-01-27 16:00 +0100

csiph-web