[PATCH v6 07/19] mm: Put readahead pages in cache earlier

[Dave Chinner]

On Mon, Feb 17, 2020 at 10:45:52AM -0800, Matthew Wilcox wrote:
> From: "Matthew Wilcox (Oracle)" <willy at infradead.org>
> 
> At allocation time, put the pages in the cache unless we're using
> ->readpages.  Add the readahead_for_each() iterator for the benefit of
> the ->readpage fallback.  This iterator supports huge pages, even though
> none of the filesystems to be converted do yet.

This could be better written - took me some time to get my head
around it and the code.

"When populating the page cache for readahead, mappings that don't
use ->readpages need to have their pages added to the page cache
before ->readpage is called. Do this insertion earlier so that the
pages can be looked up immediately prior to ->readpage calls rather
than passing them on a linked list. This early insert functionality
is also required by the upcoming ->readahead method that will
replace ->readpages.

Optimise and simplify the readpage loop by adding a
readahead_for_each() iterator to provide the pages we need to read.
This iterator also supports huge pages, even though none of the
filesystems have been converted to use them yet."

> +static inline struct page *readahead_page(struct readahead_control *rac)
> +{
> +	struct page *page;
> +
> +	if (!rac->_nr_pages)
> +		return NULL;

Hmmmm.

> +
> +	page = xa_load(&rac->mapping->i_pages, rac->_start);
> +	VM_BUG_ON_PAGE(!PageLocked(page), page);
> +	rac->_batch_count = hpage_nr_pages(page);

So we could have rac->_nr_pages = 2, and then we get an order 2
large page returned, and so rac->_batch_count = 4.
> +
> +	return page;
> +}
> +
> +static inline void readahead_next(struct readahead_control *rac)
> +{
> +	rac->_nr_pages -= rac->_batch_count;
> +	rac->_start += rac->_batch_count;

This results in rac->_nr_pages = -2 (or a huge positive number).
That means that readahead_page() will not terminate when it should,
and potentially will panic if it doesn't find the page that it
thinks should be there at rac->_start + 4...

> +#define readahead_for_each(rac, page)					\
> +	for (; (page = readahead_page(rac)); readahead_next(rac))
> +
>  /* The number of pages in this readahead block */
>  static inline unsigned int readahead_count(struct readahead_control *rac)
>  {
> diff --git a/mm/readahead.c b/mm/readahead.c
> index bdc5759000d3..9e430daae42f 100644
> --- a/mm/readahead.c
> +++ b/mm/readahead.c
> @@ -113,12 +113,11 @@ int read_cache_pages(struct address_space *mapping, struct list_head *pages,
>  
>  EXPORT_SYMBOL(read_cache_pages);
>  
> -static void read_pages(struct readahead_control *rac, struct list_head *pages,
> -		gfp_t gfp)
> +static void read_pages(struct readahead_control *rac, struct list_head *pages)
>  {
>  	const struct address_space_operations *aops = rac->mapping->a_ops;
> +	struct page *page;
>  	struct blk_plug plug;
> -	unsigned page_idx;
>  
>  	blk_start_plug(&plug);
>  
> @@ -127,19 +126,13 @@ static void read_pages(struct readahead_control *rac, struct list_head *pages,
>  				readahead_count(rac));
>  		/* Clean up the remaining pages */
>  		put_pages_list(pages);
> -		goto out;
> -	}
> -
> -	for (page_idx = 0; page_idx < readahead_count(rac); page_idx++) {
> -		struct page *page = lru_to_page(pages);
> -		list_del(&page->lru);
> -		if (!add_to_page_cache_lru(page, rac->mapping, page->index,
> -				gfp))
> +	} else {
> +		readahead_for_each(rac, page) {
>  			aops->readpage(rac->file, page);
> -		put_page(page);
> +			put_page(page);
> +		}
>  	}

Nice simplification and gets rid of the need for rac->mapping, but I
still find the aops variable weird.

> -out:
>  	blk_finish_plug(&plug);
>  }
>  
> @@ -159,6 +152,7 @@ void __do_page_cache_readahead(struct address_space *mapping,
>  	unsigned long i;
>  	loff_t isize = i_size_read(inode);
>  	gfp_t gfp_mask = readahead_gfp_mask(mapping);
> +	bool use_list = mapping->a_ops->readpages;
>  	struct readahead_control rac = {
>  		.mapping = mapping,
>  		.file = filp,

[ I do find these unstructured mixes of declarations and
initialisations dense and difficult to read.... ]

> @@ -196,8 +190,14 @@ void __do_page_cache_readahead(struct address_space *mapping,
>  		page = __page_cache_alloc(gfp_mask);
>  		if (!page)
>  			break;
> -		page->index = offset;
> -		list_add(&page->lru, &page_pool);
> +		if (use_list) {
> +			page->index = offset;
> +			list_add(&page->lru, &page_pool);
> +		} else if (add_to_page_cache_lru(page, mapping, offset,
> +					gfp_mask) < 0) {
> +			put_page(page);
> +			goto read;
> +		}

Ok, so that's why you put read code at the end of the loop. To turn
the code into spaghetti :/

How much does this simplify down when we get rid of ->readpages and
can restructure the loop? This really seems like you're trying to
flatten two nested loops into one by the use of goto....

Cheers,

Dave.
-- 
Dave Chinner
david at fromorbit.com