Received-SPF: pass (google.com: domain of linux-kernel+bounces-47892-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.48.161 as permitted sender) client-ip=147.75.48.161;
Date: Thu, 1 Feb 2024 10:37:49 +0100
From: Jan Kara <jack@suse.cz>
To: Liu Shixin <liushixin2@huawei.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>,
	Christian Brauner <brauner@kernel.org>, Jan Kara <jack@suse.cz>,
	Matthew Wilcox <willy@infradead.org>,
	Andrew Morton <akpm@linux-foundation.org>,
	linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-mm@kvack.org
Subject: Re: [PATCH 2/2] mm/readahead: limit sync readahead while too many
 active refault
Message-ID: <20240201093749.ll7uzgt7ixy7kkhw@quack3>
References: <20240201100835.1626685-1-liushixin2@huawei.com>
 <20240201100835.1626685-3-liushixin2@huawei.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240201100835.1626685-3-liushixin2@huawei.com>

On Thu 01-02-24 18:08:35, Liu Shixin wrote:
> When the pagefault is not for write and the refault distance is close,
> the page will be activated directly. If there are too many such pages in
> a file, that means the pages may be reclaimed immediately.
> In such situation, there is no positive effect to read-ahead since it will
> only waste IO. So collect the number of such pages and when the number is
> too large, stop bothering with read-ahead for a while until it decreased
> automatically.
> 
> Define 'too large' as 10000 experientially, which can solves the problem
> and does not affect by the occasional active refault.
> 
> Signed-off-by: Liu Shixin <liushixin2@huawei.com>

So I'm not convinced this new logic is needed. We already have
ra->mmap_miss which gets incremented when a page fault has to read the page
(and decremented when a page fault found the page already in cache). This
should already work to detect trashing as well, shouldn't it? If it does
not, why?

								Honza

> ---
>  include/linux/fs.h      |  2 ++
>  include/linux/pagemap.h |  1 +
>  mm/filemap.c            | 16 ++++++++++++++++
>  mm/readahead.c          |  4 ++++
>  4 files changed, 23 insertions(+)
> 
> diff --git a/include/linux/fs.h b/include/linux/fs.h
> index ed5966a704951..f2a1825442f5a 100644
> --- a/include/linux/fs.h
> +++ b/include/linux/fs.h
> @@ -960,6 +960,7 @@ struct fown_struct {
>   *      the first of these pages is accessed.
>   * @ra_pages: Maximum size of a readahead request, copied from the bdi.
>   * @mmap_miss: How many mmap accesses missed in the page cache.
> + * @active_refault: Number of active page refault.
>   * @prev_pos: The last byte in the most recent read request.
>   *
>   * When this structure is passed to ->readahead(), the "most recent"
> @@ -971,6 +972,7 @@ struct file_ra_state {
>  	unsigned int async_size;
>  	unsigned int ra_pages;
>  	unsigned int mmap_miss;
> +	unsigned int active_refault;
>  	loff_t prev_pos;
>  };
>  
> diff --git a/include/linux/pagemap.h b/include/linux/pagemap.h
> index 2df35e65557d2..da9eaf985dec4 100644
> --- a/include/linux/pagemap.h
> +++ b/include/linux/pagemap.h
> @@ -1256,6 +1256,7 @@ struct readahead_control {
>  	pgoff_t _index;
>  	unsigned int _nr_pages;
>  	unsigned int _batch_count;
> +	unsigned int _active_refault;
>  	bool _workingset;
>  	unsigned long _pflags;
>  };
> diff --git a/mm/filemap.c b/mm/filemap.c
> index 750e779c23db7..4de80592ab270 100644
> --- a/mm/filemap.c
> +++ b/mm/filemap.c
> @@ -3037,6 +3037,7 @@ loff_t mapping_seek_hole_data(struct address_space *mapping, loff_t start,
>  
>  #ifdef CONFIG_MMU
>  #define MMAP_LOTSAMISS  (100)
> +#define ACTIVE_REFAULT_LIMIT	(10000)
>  /*
>   * lock_folio_maybe_drop_mmap - lock the page, possibly dropping the mmap_lock
>   * @vmf - the vm_fault for this fault.
> @@ -3142,6 +3143,18 @@ static struct file *do_sync_mmap_readahead(struct vm_fault *vmf)
>  	if (mmap_miss > MMAP_LOTSAMISS)
>  		return fpin;
>  
> +	ractl._active_refault = READ_ONCE(ra->active_refault);
> +	if (ractl._active_refault)
> +		WRITE_ONCE(ra->active_refault, --ractl._active_refault);
> +
> +	/*
> +	 * If there are a lot of refault of active pages in this file,
> +	 * that means the memory reclaim is ongoing. Stop bothering with
> +	 * read-ahead since it will only waste IO.
> +	 */
> +	if (ractl._active_refault >= ACTIVE_REFAULT_LIMIT)
> +		return fpin;
> +
>  	/*
>  	 * mmap read-around
>  	 */
> @@ -3151,6 +3164,9 @@ static struct file *do_sync_mmap_readahead(struct vm_fault *vmf)
>  	ra->async_size = ra->ra_pages / 4;
>  	ractl._index = ra->start;
>  	page_cache_ra_order(&ractl, ra, 0);
> +
> +	WRITE_ONCE(ra->active_refault, ractl._active_refault);
> +
>  	return fpin;
>  }
>  
> diff --git a/mm/readahead.c b/mm/readahead.c
> index cc4abb67eb223..d79bb70a232c4 100644
> --- a/mm/readahead.c
> +++ b/mm/readahead.c
> @@ -263,6 +263,10 @@ void page_cache_ra_unbounded(struct readahead_control *ractl,
>  			folio_set_readahead(folio);
>  		ractl->_workingset |= folio_test_workingset(folio);
>  		ractl->_nr_pages++;
> +		if (unlikely(folio_test_workingset(folio)))
> +			ractl->_active_refault++;
> +		else if (unlikely(ractl->_active_refault))
> +			ractl->_active_refault--;
>  	}
>  
>  	/*
> -- 
> 2.25.1
> 
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR